Chimpanzee Human genome differences

[someone]

On http://en.wikipedia.org/wiki/Chimpanzee_genome_project it says:

----------
Analysis of the genome was published in Nature on September 1, 2005,
in an article produced by the Chimpanzee Sequencing and Analysis
Consortium, a group of scientists which is supported in part by
the National Human Genome Research Institute, one of the National
Institutes of Health. The article marked the completion of the
draft genome sequence[2]. A database[3] now exists containing
the genetic differences between human and chimpanzee genes, with
about thirty-five million single-nucleotide changes, five million insertion/deletion events, and various chromosomal rearrangements.
Gene duplications account for most of the sequence differences
between humans and chimps. Single-base-pair substitutions account
for about half as much genetic change as does gene duplication.
----------

Perhaps I'm misunderstanding the data here, but if we imagine that
the chimp and the human are equidistant in terms of mutations from
the common ancestor, and ignore that the chimp and the human may
have undergone the same mutations (so that there would be extra
mutations which don't show up as differences), that would still
seem to be 17.5 million single nucleotide changes that had spread
through the population and 2.5 million insertion/deletion events.

Does this not sound a lot for around 5 million years separation?

[Mitchell Coffey]

No, why would it be?

Mitchell Coffey

[jillery]

Keep in mind there are approx. 3 billion (10^9) base pairs in these
genomes. So even 17.5 million SNPs represents only 0.6% of the entire
genome. So yes, it's likely that some SNPs are duplicates, but almost
all of them are distinct locations, and the numbers reported fall
within the error bars.

>Does this not sound a lot for around 5 million years separation?

It might sound like a lot, but again, there are lots of zeroes
floating around. 5 million years is 250 thousand 20-year generations.
That reduces to a mutation rate of 2.4 x 10^-8 per base pair per
generation. This is comparable in magnitude to the rate determined by
comparing genomes across human generations, of 1.1 x 10^-8, which I
found here:

http://johnhawks.net/node/16570


So the difference between the two methods is about double. Does that
count as "a lot" to you? I'm no expert, but I suspect the difference
is from the different assumptions being made.

[Arkalen]

That's, what, about 4 mutations a year ? If a generation is 20 years
that's about 80 mutations a generation ? Sounds pretty low actually; I
thought the latest estimates had us at around 100 or even 200 novel
mutations per person.

[someone]

How many naturally occurring mutations that are spreading through the human
population are thought to have developed in the last 2000 years? I had assumed
it was less than 100, and so the figures seemed surprising to me,since they
seem to suggest that 3-4 new single nucleotide mutations are likely to spread
throughout the whole population on average every year, and a new insertion
or deletion event every year is likely to spread throughout the whole
population. If the population has never been so big then the rate these
mutations which would go on to spread throughout humanity (assuming its
continued existence) would be being created could be expected to be higher
than ever.

[snip]

[alias Ernest Major]

In a population of constant size the rate of fixation is independent of
the size of the population. (Mutations are fixed faster in a smaller
population, but more mutations occur in a larger population, and these
two factor cancel.) In a shrinking population the rate of fixation is
greater than for a constant population (the extreme case is the founder
effect). In a growing population the rate of fixation is lower, as there
hasn't been time for recent mutations to be fixed, and the older
mutations represent the smaller pool originating in a smaller population.

Try restating "3-4 new single nucleotide mutations are likely to spread
throughout the whole population on average every year" as "3-4 rare
ancestral alleles go extinct every year".

--
alias Ernest Major

[Bill]

On Jun 14, 5:17 pm, someone <glenn.spi...@googlemail.com> wrote:
> On Friday, 14 June 2013 07:43:24 UTC+1, Mitchell Coffey  wrote:
> > On 6/14/2013 2:08 AM, someone wrote:
>

Remember that it's not that 3-4 new single nucleotide mutations spread
throughout the whole population every year. It's that 3-4 single
nucleotide mutations get fixed (go to 100% prevalence) each year. So
in a single year, all that happens is they happen to go from being
present in 99.9999% of the population to being present in 100% of the
population. Meanwhile, the prevalence of millions of other single
nucleotide mutations is fluctuating up and down every year, in a few
cases, by chance, getting close to 100% or 0%, and perhaps, getting
all the way there the next year. And remember virtually all of these
mutations are selectively neutral.

[Ron O]

On Jun 14, 1:08 am, someone <glenn.spi...@googlemail.com> wrote:

What do you expect to get?

In terms of difference in single nucleotide changes any two chimps are
different from each other 1 in 300 (within the same species not
between bonobos and trogs which is greater) base-pairs while any human
and any chimp are different from each other by around 1 in a 100.
That is what the numbers are saying. So what limits are you thinking
of? Some lineages seem to evolve faster, so there isn't anything
special with the numbers.

Ron Okimoto

[someone]

On Friday, 14 June 2013 08:54:46 UTC+1, jillery wrote:

As I pointed out at the beginning I might be misunderstanding the data.
I am overlooking lots of factors in estimating how many mutations humans
would have needed in the past 5 million years.

I had assumed that these were difference found in the chimpanzee genome and this
was thought to cover pretty much the chimpanzee population. That the insertion
events were found only in chimpanzees and not in humans. I had assumed that
if humans mainly had a certain variety of a particular gene in the collective genome, and the chimps did also, then it wouldn't come up as a difference. I
wasn't assuming that it was just a comparison of two individuals. So I had
assumed that the differences are commonly found in one group but not the other.
So I wasn't thinking that they were just any old changes, but changes that
had particularly spread through one of the populations but not the other.

[someone]

On Friday, 14 June 2013 12:25:15 UTC+1, alias Ernest Major wrote:
> On 14/06/2013 11:17, someone wrote:
>
> > On Friday, 14 June 2013 07:43:24 UTC+1, Mitchell Coffey wrote:
> >> On 6/14/2013 2:08 AM, someone wrote:
> >>> On http://en.wikipedia.org/wiki/Chimpanzee_genome_project it says:
> >>> ----------
> >>> Analysis of the genome was published in Nature on September 1, 2005,
> >>> in an article produced by the Chimpanzee Sequencing and Analysis
> >>> Consortium, a group of scientists which is supported in part by
> >>> the National Human Genome Research Institute, one of the National
> >>> Institutes of Health. The article marked the completion of the
> >>> draft genome sequence[2]. A database[3] now exists containing
> >>> the genetic differences between human and chimpanzee genes, with
> >>> about thirty-five million single-nucleotide changes, five million insertion/deletion events, and various chromosomal rearrangements.
> >>> Gene duplications account for most of the sequence differences
> >>> between humans and chimps. Single-base-pair substitutions account
> >>> for about half as much genetic change as does gene duplication

I'm not sure what you mean by the rate of fixation.

> --
>
> alias Ernest Major

[John Harshman]

Why, no, it doesn't. Why would you think so? Given neutral evolution
(and the vast bulk of the genome is junk) and a stable population, the
rate of evolution is the same as the mutation rate. Each human has
around 100 mutations. Say a generation is 20 years and that humans and
chimps diverged around 5 million years ago. That's 100 * 5 million / 20,
or around 25 million changes. Not bad for a back of the envelope
calculation.

[John Harshman]

No, your intuition is wrong. Spread is slower in a big population than
in a small one. This is, however, exactly balanced by the fact that
there are more mutations in a big population than in a small one. The
number of neutral fixations (what you mean by "spread") per unit time is
the same regardless of population size, and depends only on mutation
rate. However, if the population is increasing rapidly, there will be
fewer fixations than in a stable population. So I expect you're right
that there haven't been many fixations lately. The great bulk of those
changes happened before the human population started growing so fast.

[John Harshman]

Fixation is the technical term for "spread through the whole
population", i.e. "reach a frequency of 100%".

[Roger Shrubber]

Using the 100% definition, I don't think any human alleles
are fixed.

[alias Ernest Major]

I would expect that the base pairs of some stop codons are fixed, in
that a mutation that allows translation to proceed further is liable to
produce a non-working protein and mammalian cell regulation is finicky
about gene dosage. But in a sufficiently large population fixation
becomes a less clear concept.

For rarer genetic events like insertions, deletions, inversions and
translocations is sufficiently large population is bigger.

--
alias Ernest Major

[Roger Shrubber]

alias Ernest Major wrote:
> On 14/06/2013 15:28, Roger Shrubber wrote:
>> John Harshman wrote:
>>> On 6/14/13 5:42 AM, someone wrote:
>>>> On Friday, 14 June 2013 12:25:15 UTC+1, alias Ernest Major wrote:
>>>>> On 14/06/2013 11:17, someone wrote:
>>
>>
>>>>> Try restating "3-4 new single nucleotide mutations are likely to
>>>>> spread
>>>>> throughout the whole population on average every year" as "3-4 rare
>>>>> ancestral alleles go extinct every year".
>>>>>
>>>>
>>>> I'm not sure what you mean by the rate of fixation.
>>>
>>> Fixation is the technical term for "spread through the whole
>>> population", i.e. "reach a frequency of 100%".
>>
>> Using the 100% definition, I don't think any human alleles
>> are fixed.
>>
> I would expect that the base pairs of some stop codons are fixed, in
> that a mutation that allows translation to proceed further is liable to
> produce a non-working protein and mammalian cell regulation is finicky
> about gene dosage. But in a sufficiently large population fixation
> becomes a less clear concept.
>
> For rarer genetic events like insertions, deletions, inversions and
> translocations is sufficiently large population is bigger.

I believe that the formal definition of fixation does not actually
refer to the contemporaneous extant genomes. Instead a fixed allele
refers to all instances of that current gene having a single common
ancestor, even if subsequent mutations have changed the allele.
And it's the most recent common ancestor. So a fixed allele is more
of a historical notion of "surviving heirs" even if new mutations
have occurred since. It's a strange definition at first blush but
it works for formal population genetics.

[Paul J Gans]

Of course not. If they aren't broken, why would they need
fixing?

--
--- Paul J. Gans

[jillery]

Ok. I took a shot at what you were asking, and I missed. So I have
to echo the same question others asked; what is it about the number
that sounds like "a lot" to you?

[John Harshman]

Yes, there are some problems with large populations, as a mutation is
likely to happen in the fixing lineage long before the allele is fixed.
You just have to accept that fixation of that allele or its descendant
still counts. There are other problems because of the big recent
increase in human population, which should mean the fewer alleles (or
descendants) will become fixed until the population has had a chance to
reach equilibrium.

[drose...@yahoo.com]

This way of saying it is a little incorrect. The data suggests that 3-4 single nucleotide mutations occur each year somewhere in the population that is alive that year. However, the single nucleotide changes could take thousands to a couple of hundred thousand years to spread through the population.

They detect the single nucleotide changes only in the current populations of chimps and humans. If every member of the extant population has the same point mutation, then one can assume that that point variation at some time has spread through a past population. However, populations that live in the past may not have shared that change in the point variation.

You can't even be sure that populations in the near past shared that gene. Certain populations haven't mixed until the recently. So a mutation that occurred one million years ago may not have spread until the thawing of the last ice age.

The assumption seems to be that there were 3-4 point mutations occurring each year that persisted and spread through the entire population until no member of the population didn't have it. If the point mutation occurred one million years ago, it just has to persist.

The spread can be very slow. If it is a neutral mutation, then it may remain at equilibrium for a long time in one area. It would persist, though not take over. However, slightly beneficial mutation would spread very slowly relative to lifetimes. A mutation that occurs 1 million years ago could take up to a million years to spread through the population.

A subpopulation could have been isolated from the rest of the population for a long time. The final spread could have occurred much later when climate or geography changed. So there is no need for a point variation to spread in one year.

>If the population has never been so big then the rate these
> mutations which would go on to spread throughout humanity (assuming its
> continued existence) would be being created could be expected to be higher
> than ever.

The total rate of mutation would go up with population. If there was a large population, the rate per generation would be very big. The larger the rate, the longer the time period it would take to spread through the population. However, the spread is very short relative to the time available. If the point variation took ten thousand years to spread, but occurred one million years ago, then obviously it would be all through the population by now.

Note on different thread: Are you satisfied with the multicellularity story, now? Sort of? Not quite?

[someone]

On Friday, 14 June 2013 21:02:20 UTC+1, jillery wrote:

Well it was that I thought I'd read threads on this forum where creationists
were asking about examples of recent mutations that have a beneficial effect.
And I seem to remember reading some article about some bacteria found existing
outside some Japanese company's work place in a pond, where it was consuming
plastics. I remember having the impression that the company in question was
did work in the area of genetic manipulation, but I'm not sure whether it was
correct, as I don't remember what I based the impression on. It then just
seem kind of strange to me because if on average every decade, a new
insertion or deletion event had of been expected to go through the human
population such that it would be counted by researchers as a variation in
the human population, then I would have expected many of them to be beneficial
mutations. If there were that many beneficial mutations happening at that rate
in humans then I'm surprised the answers given to creationists didn't have some
examples of those rather than plastic consuming bacteria. Since I hadn't seen
those answers I was surprised at the results.

Did you think my assumptions were ok by the way?

[Arkalen]

As has been pointed out to you, fixation is a gradual and *slow* process
- a new insertion or deletion "going through the human population"
"every decade" (not sure where you're getting those numbers btw) doesn't
mean novel genes going from zero to near-100% frequence in a few
decades; it means a few of the genes that were already near-100% getting
to be pretty-much-100% every few decades. So any gene that became common
in the population wouldn't be *recent*. They'd be recent in the same
sense that the genes for lactose tolerance are recent.

We *do* have examples of novel beneficial human mutations - off the top
of my head, look up "German super-baby" or that gene that one family in
Italy has that protects against heart disease or something or other.

Note that neither of those mutations have spread through the whole human
population yet... because they're *recent* (even the Italian mutation
originated quite a few generations ago).

You are also vastly underestimating the proportion of mutations that are
neutral. Most of them are.

Finally, you are vastly overestimating the number of genomes that are
routinely sequenced. There are tons of people walking around with
chromosomal abnormalities, chimerism, and yes beneficial mutations that
will never know about it because nobody looked at their genome because
there isn't a compelling reason for their genome to be looked at. If
you're not obviously ill, or obviously different like the German
super-baby, or don't bear a child that genetic testing reveals "isn't
yours", and don't participate in a research project that involves
looking at your genes, nobody's going to find out about any novel
beneficial mutation you might have, or any other surprising genetic
stuff you have going on. This might change in as genome sequencing and
so on become cheaper and available to all but in the meantime there you are.

[drose...@yahoo.com]

Some of what you are saying is not completely clear. So I can't really say in general. I can give one or two specific cases of being wrong.

I mentioned this first one. You seem to believe that a mutated gene has to spread entirely through the population in the first year it appeared in order to be fixed. You are assuming the gene will die out unless it spreads this rapidly. As I said, this isn't true.

Beyond that, I am unsure because you were unclear. So if I misrepresent you further on, it is a mistake not a lie. However, I have to try.

One thing you have to understand is that it is not possible at the present time to determine a priori how a gene effects development. Especially in genes that have a very small effect. Mutations that have a small effect are generally not detected until generations after the mutation occurred. So it is impossible to count how many good mutations occur each year as opposed to small mutations.

Mutations that have a large effect are usually seen in the first generation. Hence, these are the mutations that are most studied. These are also the mutations that are most likely to kill the offspring.

So there is a sampling bias when studying mutations. The best studied mutations are always bad. Big mutation means bad mutation. This creates the impression that all mutations are bad. Good mutations are hard to detect. However, they may be more common then you think.

Going by your units, you imply that the rate of mutation in a population per year is constant independent of the size of the population. This also is untrue. However, there is more than one type of mutation. Let us consider two types of mutation.

The most obvious type of mutation is a the substitution of a point gene. The probability density of point gene substitution is roughly proportional to the number of individuals in the population. Most point substitutions occur during meiosis. The length of the chromosome does not change. So this one should be measured in units of probability per generation per individual.

The more difficult type of mutation to understand is the duplication of a gene on the chromosome. By sneaky, A copy of a point gene is created, that adds to the length of the chromosome. You will note that in your link, this type of mutation was the second most common mutation. However,the given rate doesn't tell the whole story.

Your definition of random may have to be extended a bit to understand gene duplication. Such duplications are random with respect to the fitness of the organism. However, they are not completely random in the sense you are talking about. Here is the problem.

Every time a gene is duplicated, the probability of that mutation occurring later increases. In fact, the probability of copies of that gene duplicating again goes up exponentially. Every copy has about the same probability of duplicating itself. So if there were no preferential deletions of that gene, the number of genes on that chromosome goes up exponentially.

Note that this type of mutation is still random with respect to survival. Bad or good, each copy has the same chance of copying itself. If this was a bad gene, causing death or infertility, then the duplicated gene will probably be eliminated by natural selection. However, if it is a good gene then each individual copy has a better chance to survive. The more genes, the more duplication. So even if the mutation is individually rare, the probability of this mutation does go up in time.

When duplicate copies of a gene are present in a cell, then one of several things could happen. One effect could be an exaggeration of the phenotype that the gene represents.

It sounds like a challenge to Darwin's original theory, since the probability of mutation is indirectly influenced by the survival value. However, individuals don't really count in the statistics of natural selection. In terms of the fitness of the individual gene, genes that duplicated themselves are more fit than the genes that don't. So what is really happening is not a "true mutation". It is reproduction of an individual gene.

Selfish gene theory is an interesting variation on Darwin's theory of evolution. This model basically saves the idea of random. Note that the copies of a particular gene can also undergo point substitution.

Duplication of genes on a chromosome is a very important type of mutation. The variation of this type of mutation can increase much faster than the other types of mutation under the right conditions.

Note that the reverse type mutation, deletion of point genes, can disappear quickly. If a gene is prone to deletion, it will vanish very fast. Especially if it is bad.

Getting back to sampling bias, gene deletion is the best studied form of mutation. This is because the effects of gene deletion are often very large. They can be seen in the first generation. Most of the famous monsters in biology books are caused by gene deletions. However, they are also the least important ones for evolution.

A gene that tends to have itself deleted will tend to vanish whether it is "good" or not. Further, gene deletions tend to be bad. However, they are the most easily studied.

Gene deletions are the best known mutations. However, I suggest that you read up on other forms of mutation. Gene duplication, chromosome rearrangement, and chromosome fusion and genome duplication are far more important to evolution. These have very small effects on the phenotype. There is some experimental work concerning these mutations.

Good mutations have been studied. However, they generally are detected generations after they occurred. A mutation to a fruit fly could be good. However, it may not be detected until it has spread through the entire population. So then it is impossible to say when it occurred, exactly.

In order to be absolutely sure when a mutation occurred, one has to map the entire genome of each individual in a population every generation. This is hard in the case of simple organisms and impossible in human beings.


The other thing is that a mutation can have more than one effect. One effect can be good and one bad. However, the good effect can outdo the bad effect. The best known genes with dual fitness are the sickle cell anemia and Tay Sachs disease. I think that this may be common to most genes. Even a good gene will have a bad effect. So the fraction of good versus bad genes varies with environment. So how would a scientist be able to determine good versus bad genes in general?

If you reply, then please clip without mercy. I am sworn not to clip. So you have to do it.

[Rolf]

I like reading what you write but double spacing is worse even than looong
lines.
Rolf

<drose...@yahoo.com> skrev i melding
news:44ce23ef-8f7e-4504...@googlegroups.com...

[Rolf]

"Arkalen" <ark...@inbox.com> skrev i melding
news:Jym*lr...@news.chiark.greenend.org.uk...

IIRC, it was something to do with susceptibility to clotting of arteries by
cholesterol.

I also think there are people with resistance against the HIV virus.
I just googled 'hiv resistance' and got
http://www.wired.com/medtech/health/news/2005/01/66198?currentPage=all

[eridanus]

I had never studied this matter of genetics, but what you said in this
post do not look right to me.

I suppose you mean these mutations do not spread suddenly over the
nearly 100 % population in a single year, for it would be a miracle.
But keeps accumulating among a population for years, till they had
spread to almost the whole population. Then, I am not an expert in these
matters, Ron Okimoto would tell more. But I think that genetic changes
either spread slowly year from year, for the children that inherit them
are becoming more numerous, and it would take a long period of time
until it is considered nearly 100% in a population.

But in small groups, changes can occur much faster. That is the reason
why species isolated in islands or in forest, changed so fast compared
to great populations.

Let's us see what Ron says.

Eridanus

[jillery]

Ernest Major already explained it in this topic:

"In a population of constant size the rate of fixation is independent
of the size of the population. (Mutations are fixed faster in a
smaller population, but more mutations occur in a larger population,
and these two factor cancel."

Yes, once a mutation happens, it will tend to fix faster in a smaller
populations than in a larger population. But mutations don't happen
as often in small populations. Mathematically and in reality, the
faster fixation in smaller populations exactly cancels the greater
numbers of mutations in larger populations, making the rate of
fixation, ie the number of fixed mutations divided by the total number
of mutations, independent of population size, as Ernest wrote.

[someone]

I was assuming that neutral mutations wouldn't particularly spread through
the population, and that spread through a population would rely on beneficial
mutations. Therefore I was thinking that a significant proportion of the 35
million single nucleotide changes and 5 million insertion or deletion events
which as I understand it genetically distinguish between the human and
chimpanzee life forms and have developed in the last 5 million years, would be
beneficial. So I didn't think they were talking about any old neutral mutations.

Thanks for the references to the mutations. In the article
http://singularityhub.com/2009/06/30/super-babies-reveal-the-key-to-strength-gene/

It seems strange that there were no super babies born before and then in
2004 and 2005 there appears one in Germany and then the USA, but there you
go. The report seems to suggest that this is the degradation of a mutation
that was presumably beneficial (the production of myostatin). Also looking
at the pictures of cattle, I guess I'm slightly surprised that this didn't
happen in more animal species as it would seem to be advantageous for
species that rely on physical combat for mating rights. Though wild animals
in environments where there are seasonal food shortages would have consumed
more calories presumably because of their muscle mass. Though on the other hand,
once enough food wasn't consumed to support such a muscle mass the food taken
in would stop being put towards muscle generation and the muscle mass would
start to be used as a food source instead. If the latter consideration
outweighed the former, then it would seem to offer some species a competitive
advantage when food is in greater supply, and also act as an energy bank when
it isn't. Assuming Myostatin isn't vitally involved in the cessation of muscle
mass development in times of food shortage.

The article that I saw on the Italian Gene
http://articles.latimes.com/1994-10-17/news/mn-51362_1_artery-disease

Didn't really explain too much about the mutation, and so I can't tell
whether it was a degradation of a mutational feature that was presumably
beneficial in earlier human history or whether it would have been an
enhancement at any point in the prior human evolutionary history.

[Bill]

On Jun 17, 6:38 pm, someone <glenn.spi...@googlemail.com> wrote:

>
> I was assuming that neutral mutations wouldn't particularly spread through
> the population, and that spread through a population would rely on beneficial
> mutations. Therefore I was thinking that a significant proportion of the 35
> million single nucleotide changes and 5 million insertion or deletion events
> which as I understand it genetically distinguish between the human and
> chimpanzee life forms and have developed in the last 5 million years, would be
> beneficial. So I didn't think they were talking about any old neutral mutations.

Ah, but it was "any old neutral mutation" that they were talking
about, and they do indeed spread through the population by random
genetic drift. You can read about it in wikipedia

http://en.wikipedia.org/wiki/Genetic_drift

Ernest Major's original answer to you gave the highlights; the wiki
article just fills in all the details.


>
> Thanks for the references to the mutations. In the articlehttp://singularityhub.com/2009/06/30/super-babies-reveal-the-key-to-s...

>
> It seems strange that there were no super babies born before and then in
> 2004 and 2005 there appears one in Germany and then the USA, but there you
> go. The report seems to suggest that this is the degradation of a mutation
> that was presumably beneficial (the production of myostatin). Also looking
> at the pictures of cattle, I guess I'm slightly surprised that this didn't
> happen in more animal species as it would  seem to be advantageous for
> species that rely on physical combat for mating rights. Though wild animals
> in environments  where there are seasonal food shortages would have consumed
> more calories presumably because of their muscle mass. Though on the other hand,
> once enough food wasn't consumed to support such a muscle mass the food taken
> in would stop being put towards muscle generation and the muscle mass would
> start to be used as a food source instead. If the latter consideration
> outweighed the former, then it would seem to offer some species a competitive
> advantage when food is in greater supply, and also act as an energy bank when
> it isn't. Assuming Myostatin isn't vitally involved in the cessation of muscle
> mass development in times of food shortage.
>

> The article that I saw on the Italian Genehttp://articles.latimes.com/1994-10-17/news/mn-51362_1_artery-disease

[someone]

I'm still not clear on which way you are using it, your reference to
ancestral alleles hints that you might be using it as the poster
Roger Shrubber suggested "a fixed allele refers to all instances of
that current gene having a single common ancestor", or you might
mean be using the 100% definition, as it seems plausible that
some stop codons (I assume from context that these are some chemical
arrangements that signal the end of the protein production) are through
a 100% of the population and that you supplied an example where the rarer
case was plausible could if not a coincidence indicate that you are using
the 100% definition.


> --
>
> alias Ernest Major

[chris thompson]

On Jun 14, 10:28 am, Roger Shrubber <rog.shrub...@gmail.com> wrote:
> John Harshman wrote:
> > On 6/14/13 5:42 AM, someone wrote:
> >> On Friday, 14 June 2013 12:25:15 UTC+1, alias Ernest Major  wrote:
> >>> On 14/06/2013 11:17, someone wrote:
> >>> Try restating "3-4 new single nucleotide mutations are likely to spread
> >>> throughout the whole population on average every year" as "3-4 rare
> >>> ancestral alleles go extinct every year".
>
> >> I'm not sure what you mean by the rate of fixation.
>
> > Fixation is the technical term for "spread through the whole
> > population", i.e. "reach a frequency of 100%".
>
> Using the 100% definition, I don't think any human alleles
> are fixed.

Perhaps not in nuclear DNA. But in mitochondrial DNA I believe most of
the alleles are fixed. If you have a mutation in your mtDNA, (and some
people do) you are totally screwed. Google "mitochondrial disease"

Chris

[Bill]

On Jun 17, 6:54 pm, someone <glenn.spi...@googlemail.com> wrote:
> On Friday, 14 June 2013 15:55:06 UTC+1, alias Ernest Major  wrote:

>
>
>
>
>
>
>
>
>
> > On 14/06/2013 15:28, Roger Shrubber wrote:
> > > John Harshman wrote:
> > >> On 6/14/13 5:42 AM, someone wrote:
> > >>> On Friday, 14 June 2013 12:25:15 UTC+1, alias Ernest Major  wrote:
> > >>>> On 14/06/2013 11:17, someone wrote:
> > >>>> Try restating "3-4 new single nucleotide mutations are likely to spread
> > >>>> throughout the whole population on average every year" as "3-4 rare
> > >>>> ancestral alleles go extinct every year".
> > >>> I'm not sure what you mean by the rate of fixation.
> > >> Fixation is the technical term for "spread through the whole
> > >> population", i.e. "reach a frequency of 100%".
> > > Using the 100% definition, I don't think any human alleles
> > > are fixed.

> > I would expect that the base pairs of some stop codons are fixed, in
> > that a mutation that allows translation to proceed further is liable to
> > produce a non-working protein and mammalian cell regulation is finicky
> > about gene dosage. But in a sufficiently large population fixation
> > becomes a less clear concept.
>
> > For rarer genetic events like insertions, deletions, inversions and
> > translocations is sufficiently large population is bigger.
>
> I'm still not clear on which way you are using it, your reference to
> ancestral alleles hints that you might be using it as the poster
> Roger Shrubber suggested "a fixed allele refers to all instances of
> that current gene having a single common ancestor", or you might
> mean be using the 100% definition, as it seems plausible that
> some stop codons (I assume from context that these are some chemical
> arrangements that signal the end of the protein production) are through
> a 100% of the population and that you supplied an example where the rarer
> case was plausible could if not a coincidence indicate that you are using
> the 100% definition.

Just think about how you measure fixation. You have a human population
of several billion. You won't test them all, so the best you'll ever
be able to say is that all samples tested have the same allele. Based
on the number of samples you test you can estimate a lower limit on
the frequency of the allele, and if you sample enough, that lower
limit will be very close to 100%, but that's the best you can do.

[John Harshman]

Depends on the mutation. Many mutations, even in mitochondrial DNA,
don't matter.

[Arkalen]

Again, your assumption was incorrect. That's fine, I thought the maths
of fixation were counter-intuitive when I first learned about them, but
it might be quicker if you read the Wikipedia page for "fixation" to get
the basics and came back to us with any questions you have left ?

The thing about the way alleles spread in a population is that it's
basically a random walk : the frequency of an allele in the population
will go up or down every generation depending on the vagaries of
reproduction. The thing is, if you start out with a certain set of
alleles that are all as likely to occur in the next generation (i.e.
they're all neutral mutations with respect to each other), eventually
after enough generations *only one allele will be left*. Simply because
those frequencies will randomly fluctuate between 0% and 100%, and the
second one of them reaches 0% it stays there, so eventually one reaches
100% and that allele is fixed in the population.

That happens whether the alleles are neutral, beneficial or harmful. The
only difference is that beneficial mutations are *more likely* to become
fixed (the more beneficial they are, the higher the likelihood), and are
likely to become fixed after fewer generations, and vice-versa for
harmful mutations.

Note also that all living organisms are very close to optimized for
their environment, so that doesn't only mean mutations are more likely
to be harmful than beneficial (though they're most likely to be
neutral), but beneficial mutations will tend to be slightly beneficial,
while harmful mutations easily range from mildly harmful to lethal.

In other words, neutral mutations spread through the population not
quite but almost as easily as slightly-beneficial mutations do. And most
beneficial mutations are slightly-beneficial.

> Therefore I was thinking that a significant proportion of the 35
> million single nucleotide changes and 5 million insertion or deletion events
> which as I understand it genetically distinguish between the human and
> chimpanzee life forms and have developed in the last 5 million years, would be
> beneficial. So I didn't think they were talking about any old neutral mutations.
>
> Thanks for the references to the mutations. In the article
> http://singularityhub.com/2009/06/30/super-babies-reveal-the-key-to-strength-gene/
>
> It seems strange that there were no super babies born before and then in
> 2004 and 2005 there appears one in Germany and then the USA, but there you
> go. The report seems to suggest that this is the degradation of a mutation
> that was presumably beneficial (the production of myostatin). Also looking
> at the pictures of cattle, I guess I'm slightly surprised that this didn't
> happen in more animal species as it would seem to be advantageous for
> species that rely on physical combat for mating rights. Though wild animals
> in environments where there are seasonal food shortages would have consumed
> more calories presumably because of their muscle mass. Though on the other hand,
> once enough food wasn't consumed to support such a muscle mass the food taken
> in would stop being put towards muscle generation and the muscle mass would
> start to be used as a food source instead. If the latter consideration
> outweighed the former, then it would seem to offer some species a competitive
> advantage when food is in greater supply, and also act as an energy bank when
> it isn't. Assuming Myostatin isn't vitally involved in the cessation of muscle
> mass development in times of food shortage.

Note that humans are particular in that a lot of our evolution has
involved the loss of muscles, such as jaw muscles and the kind of
muscles that make chimpanzees ridiculously strong compared to us. So a
mutation restoring some of those muscles and dramatically increasing
strength is a low-hanging fruit so to speak. Species that are selected
for physical strength already *have* all those muscles; any mutation
making them stronger will be much more incremental.

Kind of like it's fairly easy for few mutations in a snake to result in
legs, but you'll never see that in a salmon.

And then there is also the issue that you point out, that it's all about
tradeoffs. More muscles can be advantageous in some situations and
disadvantageous in others. Which kind of illustrates how pointless it is
to say that all mutations are harmful, or that a beneficial mutation
doesn't count because in some other circumstance it would have been harmful.

>
> The article that I saw on the Italian Gene
> http://articles.latimes.com/1994-10-17/news/mn-51362_1_artery-disease
>
> Didn't really explain too much about the mutation, and so I can't tell
> whether it was a degradation of a mutational feature that was presumably
> beneficial in earlier human history or whether it would have been an
> enhancement at any point in the prior human evolutionary history.
>

It's a pretty nifty enhancement now, that's for sure. So does
"beneficial" to you mean "beneficial in every possible circumstance", or
"at all points in history" ?

[someone]

On Monday, June 17, 2013 12:45:47 PM UTC+1, Bill wrote:

> On Jun 17, 6:38 pm, someone wrote:
> > I was assuming that neutral mutations wouldn't particularly spread through
> > the population, and that spread through a population would rely on beneficial
> > mutations. Therefore I was thinking that a significant proportion of the 35
> > million single nucleotide changes and 5 million insertion or deletion events
> > which as I understand it genetically distinguish between the human and
> > chimpanzee life forms and have developed in the last 5 million years, would be
> > beneficial. So I didn't think they were talking about any old neutral mutations.
>
> Ah, but it was "any old neutral mutation" that they were talking
> about, and they do indeed spread through the population by random
> genetic drift. You can read about it in wikipedia
>
> http://en.wikipedia.org/wiki/Genetic_drift
>
> Ernest Major's original answer to you gave the highlights; the wiki
> article just fills in all the details.
>

I read the article and it was useful, though I'm not exactly clear on
what you are suggesting in the article supports the idea that
we should expect any old mutations to spread through the population
by random genetic drift.

The article does give what seems to be intended to be an example of
how neutral alleles could spread through the population. The scenario
starts like this:
----------
The mechanisms of genetic drift can be illustrated with a simplified
example. Consider a very large colony of bacteria isolated in a drop
of solution. The bacteria are genetically identical except for a
single gene with two alleles labeled A and B. Half the bacteria have
allele A and the other half have allele B. Thus both A and B have allele
frequency 1/2. A and B are neutral alleles—meaning they do not affect
the bacteria's ability to survive and reproduce. This being the case,
all bacteria in this colony are equally likely to survive and reproduce.
The drop of solution then shrinks until it has only enough food to sustain
four bacteria. All the others die without reproducing. Among the four who
survive, there are sixteen possible combinations for the A and B alleles
----------

Then, as I understand it, the author(s) go onto consider that out of the
enormous population only 4 survive and then consider the that 50% of the time
one of the alleles will become 4 times more frequent than the other in the
example. This mechanism for one allele become much more frequent than the other
even when both are neutral they refer to as "population bottleneck" in the
example the population shrunk to 4.

The reason I don't think that the article supports the idea of supports
the idea that we should expect any old mutations to spread through the
population by random genetic drift is that while they managed to get it
to 10/16 times a large shift in allele distribution would occur the example
they used was at an extreme end of the spectrum.

For example if you were use their example but to consider a population
bottle neck of 9 individuals (you can use Pascal's Triangle)
--------------------------
A | B | Combinations
--------------------------
0 | 9 | 1
1 | 8 | 9
2 | 7 | 36
3 | 6 | 84
4 | 5 |126
5 | 4 |126
6 | 3 | 84
7 | 2 | 36
8 | 1 | 9
9 | 0 | 1

You can see that chance that the resultant population
will have 3.5 times or more of one gene than the other
has dropped to 92/512 which is well under 50%, and the
trend continues the larger the population size at the
bottle neck.

Furthermore in the example there were only two types of
genes and they had already spread through the population
to 50% each. It changes considerably when the mutation is
just in one individual. As they later say (I am assuming
they are correct):
----------
Assuming genetic drift is the only evolutionary force
acting on an allele, at any given time the probability
that an allele will eventually become fixed in the
population is simply its frequency in the population at
that time.[16] For example, if the frequency p for allele
A is 75% and the frequency q for allele B is 25%, then
given unlimited time the probability A will ultimately
become fixed in the population is 75% and the probability
that B will become fixed is 25%.
----------

So for a population of say 5000, and unlimited time,
the chances of the gene spreading through the population
is 1/5000 (according to the author(s)). If you then
consider the amount of mutations difference we are talking
about between a human and a chimpanzee (the figures are
in the millions) there would I think have had to of been
an expectation of quite a few "any old neutral mutations"
for millions of them to have been expected to have spread
throughout the population in 5 million years.
by genetic drift.

Thanks for the link to the article.

[snip]

[someone]

As the article on fixation that you pointed me to points out, at a given point
in time the chances that the mutation will ever be fixed is its frequency
in the population. So if we were to talk about a population of 5000 that
had separated off from another group, if one were to have a mutation that
was neutral, there would be a 1 in 5000 chance of it fixing in the group
(assuming the random walk is the only force acting on it). So for every
neutral mutation that fixed, you would have expected 4999 not
to of. Early population figures higher than 5000 would make it even more
unlikely for any neutral mutation to ever get fixed.

both are beneficial but degradations in mutational features that
were presumably beneficial in earlier human life are presumably more
common, and if they turn out to be regressions to earlier functionality,
then they are not the only type beneficial mutation that would have
been required.

[Greg Guarino]

On 6/17/2013 4:51 PM, someone wrote:

> So for every
> neutral mutation that fixed, you would have expected 4999 not

Exactly so. Now ask yourself how many (new) mutations we would expect to
find in a population of 5000.

[Bill]

On Jun 18, 3:20 am, someone <glenn.spi...@googlemail.com> wrote:

>
> > Ah, but it was "any old neutral mutation" that they were talking
> > about, and they do indeed spread through the population by random
> > genetic drift. You can read about it in wikipedia
>
> >http://en.wikipedia.org/wiki/Genetic_drift
>
> > Ernest Major's original answer to you gave the highlights; the wiki
> > article just fills in all the details.
>
> I read the article and it was useful, though I'm not exactly clear on
> what you are suggesting in the article supports the idea that
> we should expect any old mutations to spread through the population
> by random genetic drift.

If you are not clear as to what in the article supports the idea that

we should expect "any old mutations" to spread through the population

by random genetic drift, then I fear you did not understand the
article. I'll see if I can figure out what's bothering you below.

Perhaps you have trouble imagining how a newly arising mutation could
ever spread throughout a large population. Imagine a human population
of a billion individuals. If a new, neutral mutation appears in a
single individual in that population, then the odds that that new
allele will ultimately become fixed in the population is 1/1000000000.
Rather a small chance. That's what's bothering you, I guess. But let's
do some more numbers.

The estimated chance that an individual nucleotide site will be
mutated is estimated at about 1 x 10^-8 per generation in humans. The
human genome contains about 3 x 10^9 nucleotide sites, so that gives
an estimate of about 30 single nucleotide mutations per individual.
Other estimates made by direct genomic sequencing are a bit higher,
maybe 100-200 mutations per individual. Now, in a population of a
billion humans, there will therefore be between 30-200 billion new
mutations each generation in the whole population. Of course, each
individual one of these will have only a 1 in a billion chance of
ultimately going to fixation in the population, but since there are
30-200 billion entrants in this "fixation lottery" every generation,
you expect 30-200 mutations to go to fixation each generation. Of
course it takes a long time in a huge population, but remember that
for most of the time that humans were splitting off from chimps the
human population was a whole lot smaller than 1 billion. And as you'll
understand if you reread the wiki article carefully, the number of
neutral mutations that go to fixation in each generation is a function
only of the mutation rate and is independent of the population size
(that shouldn't be hard to understand - in a huge population it's very
unlikely that any single mutation will go to fixation, but in a huge
population there are a huge number of new mutations in the population
in each generation, so a few of them win the fixation lottery).

[drose...@yahoo.com]

On Monday, June 17, 2013 4:20:52 PM UTC-4, someone wrote:
> On Monday, June 17, 2013 12:45:47 PM UTC+1, Bill wrote:
>

> frequency 1/2. A and B are neutral alleles�meaning they do not affect

According to the Hardy-Weinberg principle, without mutation or
selection the proportion of each gene remains constant for a large
population. So without selection, the fraction of single point
substitutions increases linearly with time in a large population.
So if we are talking about neutral mutations that are single point
substitutions, it seems unlikely that any single point substitution
will "take over" a population.

The situation is a little different with gene duplications. Suppose
that there is an allele that has a higher probability of duplication
than the alternate allele at the same locus. The gene duplicates and
stays on the chromosome. Assume that the gene is otherwise neutral with
respect to the survival of an individual.

If an allele has a high probability of duplicating itself, all
copies of that allele have an equally high probability of duplicating
itself. So the proportion of that gene increases exponentially in time
even if that gene is "neutral".

The paradox is resolved by refining the definition of "neutral".
Lets consider the "gene centric" fitness.

An allele that duplicates itself is more fit than an allele that
doesn't duplicate itself, even if it doesn't increase the probability
of its "host" surviving. A gene that duplicates itself could even be
called a "selfish gene", because it spreads through the population
without necessarily helping the host.

Gene duplication is a type of mutation that intrinsically increases
the information available to a cell. After a gene duplication, the genome
is just slightly more complex than before.

Please note that your article pointed out that gene duplications
were the second most common mutation inferred from the differences
between chimpanzee and human.

Your calculation only applied to point gene substitutions. Not only
that, you assumed that the mutated gene was absolutely neutral with
respect to survival. I seriously doubt that those genes which fixed
themselves were totally neutral.

For all you know, one mutation in four may be "beneficial". This is not
true for mutations that cause large effects. However, that is because a
large effect automatically extends beyond physical thresholds. If there
are only two directions to go in any feature, and the mutation is very
small, any mutation at that site has a 50% chance of being in the "right"
direction.

Also note that there is a synergy between genes. One mutated gene may
be neutral and so remain stationary in the population for a long time.
However,it may become useful when another mutation occurs.

Few popular scientists point this out. However, there is also a little
matter concerning migration.

The usual calculation of probability for a "good" gene may be small if
the offspring remains in the same environment all its life. However, organisms
move. A small mutation that is unfit in one habitat may be fit if it
moves just a small distance to another habitat.

Suppose an animal lives on the side of a mountain in the middle, for
example. Suppose that a gene duplication makes the animal furrier. If may
feel too hot if it stays at the middle of the mountain. However, the
temperature is colder as one climbs the mountain. Therefore, the animal
just has to move a small distance upward to be comfortable.

A lot of clines and ring species are like that. Small changes in the
genome of an animal are distributed over large geographic distances. The
difference between animals on one end of the cline and the other may be
quite big. Each generation adapted to a slightly shifted geographical
location.

Migration can make it very likely that a small mutation is beneficial.
If the animal or plant moves a very small distance each mutation, then
the accumulated changes due to a series of mutations can be very large
without killing the animal.

So I think an error in your assumptions, and some of the people who
are replying, is the assumption that the probability is insignificant that
a small mutation is beneficial. I think it is probably more likely than
you think.

I have heard talks by paleontologists who conclude that many types of
organisms originate in one location and move to another location, leaving
species that branch off the entire way. Usually, the first species starts
in the tropics and moves toward the poles. There is still controversy as
to why most taxons originate in the tropics, near the equator.

Could you clip most of your reply near the top?

[someone]

On Monday, June 17, 2013 10:18:33 PM UTC+1, Bill wrote:

> On Jun 18, 3:20�am, someone wrote:
>
> > > Ah, but it was "any old neutral mutation" that they were talking
> > > about, and they do indeed spread through the population by random
> > > genetic drift. You can read about it in wikipedia
> > >http://en.wikipedia.org/wiki/Genetic_drift
> > > Ernest Major's original answer to you gave the highlights; the wiki
> > > article just fills in all the details.
> > I read the article and it was useful, though I'm not exactly clear on
> > what you are suggesting in the article supports the idea that
> > we should expect any old mutations to spread through the population
> > by random genetic drift.
> If you are not clear as to what in the article supports the idea that
> we should expect "any old mutations" to spread through the population
> by random genetic drift, then I fear you did not understand the
> article. I'll see if I can figure out what's bothering you below.
>
> > The article does give what seems to be intended to be an example of
> > how neutral alleles could spread through the population. The scenario
> > starts like this:
> > ----------
> > The mechanisms of genetic drift can be illustrated with a simplified
> > example. Consider a very large colony of bacteria isolated in a drop
> > of solution. The bacteria are genetically identical except for a
> > single gene with two alleles labeled A and B. Half the bacteria have
> > allele A and the other half have allele B. Thus both A and B have allele

> > frequency 1/2. A and B are neutral alleles�meaning they do not affect

> > the bacteria's ability to survive and reproduce. This being the case,
> > all bacteria in this colony are equally likely to survive and reproduce.
> > The drop of solution then shrinks until it has only enough food to sustain
> > four bacteria. All the others die without reproducing. Among the four who
> > survive, there are sixteen possible combinations for the A and B alleles
> > ----------
>
> > Then, as I understand it, the author(s) go onto consider that out of the
> > enormous population only 4 survive and then consider the that 50% of the time
> > one of the alleles will become 4 times more frequent than the other in the
> > example. This mechanism for one allele become much more frequent than the other
> > even when both are neutral they refer to as "population bottleneck" in the
> > example the population shrunk to 4.
> > The reason I don't think that the article supports the idea of supports
> > the idea that we should expect any old mutations to spread through the
> > population by random genetic drift is that while they managed to get it
> > to 10/16 times a large shift in allele distribution would occur the example
> > they used was at an extreme end of the spectrum.
>
> > For example if you were use their example but to consider a population
> > bottle neck of 9 individuals (you can use Pascal's Triangle)
> > --------------------------

> > � A �| �B �| Combinations
> > --------------------------
> > � 0 �| �9 �| �1
> > � 1 �| �8 �| �9
> > � 2 �| �7 �| 36
> > � 3 �| �6 �| 84
> > � 4 �| �5 �|126
> > � 5 �| �4 �|126
> > � 6 �| �3 �| 84
> > � 7 �| �2 �| 36
> > � 8 �| �1 �| �9
> > � 9 �| �0 �| �1

I can see your point, though there does still seem to be a potential
issue of time. Beneficial mutations could be expected to spread through
faster than neutral mutations. If we imagine a mutation rate of 200 per person per generation and assume they are all neutral, and imagine a population of
100,000 where each couple mates for life, and has two children which get to
reproduce (so the population size can remain pretty constant), then what I'm
not clear on is how many generations would it be expected to take before a
single neutral mutation from the 20,000,000 that had occurred in the original
generation got fixed?

[snip]

[someone]

On Tuesday, June 18, 2013 8:57:29 AM UTC+1, someone wrote:
> On Monday, June 17, 2013 10:18:33 PM UTC+1, Bill wrote:
>

> > On Jun 18, 3:20�am, someone wrote:
>

[snip]

I just made an estimate myself of about 25 generations, so if my working was
ok then there doesn't seem to be a problem with the amount of time either.

There is just the issue of the expected rate of beneficial mutation that isn't
simply a regression to an earlier state. How many of the mutational differences
between a chimpanzee and human are thought to be of that type?

[Arkalen]

Indeed, but that changes the question. I was assuming you were wondering
about examples of beneficial mutations arising nowadays, to justify the
contention that beneficial mutations arise in general at a nonzero rate.
By definition any example of such a mutation will be beneficial *to us*,
*now*. They're relevant to the question of whether beneficial mutations
arise in general, but they're not necessarily good examples of
beneficial mutations that would have been useful to our proto-human
ancestors and that contributed to them becoming us.

If *that* is what you are asking about, then you're asking a historical
question, not a general one, is that correct ? To answer it we shouldn't
look at current examples of mutations, but at mutations that happened in
the past.

In particular, if you look at the specific genetic differences there are
between humans and chimpanzees (there's a Wikipedia page somewhere
mentioning them) you'll see a number of them that are thought to
correspond to physical changes that make us us - reduction in jaw
muscles, changes in the brain, etc.

Which of course doesn't change that most of the individual differences
are neutral mutations occurring in inactive parts of the genome (like
the vitamin C pseudogene).

I hope this answers your questions but I'm really losing track of how
the current discussion relates to your original question - could you
remind me what that was, and which specific aspects of it remain
unanswered in your mind ? (and maybe which you've changed your mind on
and have new questions about now ?)

[Arkalen]

On 17/06/13 21:51, someone wrote:

Sorry, I missed that first paragraph in my original reply.

Just like with the previous reply though I'm confused as to what it is
you're asking at this point. But I'm going to go on illustrating why 1)
lots of mutations do get fixed (or at 99% at least) and 2) neutral
mutations get fixed almost as often as slightly beneficial mutations.

First, let's take your population of 5000 people. It's true if a
mutation has a frequency of 1 in 5000 it has a very small probability of
getting fixed. That basically corresponds to a situation of a novel
mutation that's just appeared. It's in only one person; that person
might die before reproducing. It only has a 50% chance of passing that
mutation on to every child if it doesn't. If few children have it, they
might die or not pass it on either.

Note that these things are true *whether or not* the mutation is neutral
or slightly beneficial. If the mutation is beneficial those things are
*slightly less likely* to happen, but when you're looking at such small
numbers then it doesn't help that much.

So that novel neutral mutation has a 1 in 5000 thousand chance of ending
up, after many many generations, in the whole population. If it were
slightly beneficial, that chance might be 1.1 in 5000. Not a huge
difference is it ?

Now remember that the rate of mutation per nucleotide is low, but we
have A LOT of nucleotides. Let's say every person has 100 novel
mutations (which is close enough to the actual number IIRC).

That means every generation, in that population of 5000 people, you've
got 500,000 novel mutations.

Let's say 10% of them are beneficial, with a 10% advantage over neutral
mutations. I'll ignore harmful mutations to keep things simpler, but to
avoid overestimating the numbers let's cut out 20% of the mutations
wholesale; we'll say they're the lethal mutations, and the mildly
harmful ones that didn't happen to stick around.

So we have 400,000 novel mutations; 50,000 of them have a 1,1 chance in
5000 of ending up fixed after many generations, while 350,000 of them
have a 1 in 5000 chance of doing so.

After many generations that means we can expect 80 neutral mutations to
have become fixed, and 11 beneficial mutations to have done so. Which
conveniently is pretty close to the rate of mutation in an individual
(*because* the probability of getting fixed is the same as the frequency
in the population; if I hadn't had beneficial and harmful mutations in
there it would be exactly the same)

Notice how many more neutral mutations there are than beneficial ones,
though the beneficial mutations outperformed them (the proportion of
beneficial mutations in the first generation was 10%, their proportion
among the fixed mutations from that generation is 12%)

[Bill]

Your estimate is much too low. If you go back to the wiki article I
suggested you'll see that the expected time to fixation of a new
mutation (assuming it wins the "fixation lottery" and does ultimately
go to fixation) is approximately 4 x N generations, where N is the
effective population size. So, in a population of 100,000, you'd
expect it to take 400,000 generations, or maybe 8 million years for
fixation to occur. The time may be significantly shorter if the
effective population size is significantly smaller than the real
population size (effective population size takes into account things
like inbreeding, fluctuation in population through time, overlapping
generations, polygamy). It is often the case that effective population
size is much less than the census population. So you might expect a
few million years for a fixation to occur. And remember that the
different fixations occur in parallel, not one after the other. SO
there should be plenty of time to accumulate lots of neutral mutations
since we diverged from chimps.

[Arkalen]

I don't think we have a percentage because we don't know what all the
mutational differences *do*. There are some that are known to be neutral
(like changes in pseudogenes and ERVs), others are known or thought to
be important (changes in muscle genes and brain regulatory regions IIRC)
but I don't know if we know enough to have a proportion.

What I want to point out is that you can't rule out "simply a regression
to an earlier state", doing so ignores what modern organisms *are*.

Basically, such a distinction between "regression" and "advancement"
might have made sense in very early and simple genomes when there simply
wasn't much evolutionary history behind them (I don't know if that's
actually true; it might always have been a muddle). But modern organisms
are the result of billions of years of evolution involving additions and
subtractions, our development involves tons of different processes
pushing and pulling in different directions and regulatory systems
controlling what the final result looks like, with tons of genes
upregulating and downregulating, activating and deactivating each other
at specific times and places.

*Any* change you make in that system could be described as a
"regression" in some way or another; activating one thing will often
mean deactivating another and you can always find something in the past
the new result would be like.

In fact a lot of the differences between humans and chimpanzees involve
neoteny, i.e. the persistence of infantile traits into adulthood. You
could say that walking upright was a regression by our ancestors, a case
of arrested development.

Let's imagine that our larger brains were the result of a mutation in a
regulatory gene that leads it to inactivate a gene that would usually
downregulate genes that promote brain development in the fetal stage.
Would this be an advancement or a regression ? A loss or a gain of
information ? A deactivation or an activation ?

[alias Ernest Major]

On 18/06/2013 12:45, Arkalen wrote:
> I don't think we have a percentage because we don't know what all the

> mutational differences*do*. There are some that are known to be neutral

> (like changes in pseudogenes and ERVs), others are known or thought to
> be important (changes in muscle genes and brain regulatory regions IIRC)
> but I don't know if we know enough to have a proportion.

When a selective sweep occurs (that is when one allele is replaced by
another because of selection) the population diversity of the
neighbouring region of the genome is reduced as one haplotype hitchkikes
alone with the locus under selection. (How big a section of the genome
depends on the intensity of the selection; the more intense the
selection the bigger the chunk of the genome that is fixed before
recombination breaks the links between the locus under selection and its
neighbours.) Thus we can identify regions of our genome that have
recently undergone directional selection. After the selective sweep is
complete diversity rebuilds from a combination of mutation and drift, so
eventually the evidence of selection is erased, limiting the age of
selective events that can be recognised in this way.

Over 1000 human genomes have been sequence, so it seems to me that we
should have the information to place a lower bound on the proportion.

--
alias Ernest Major

[someone]

On Tuesday, June 18, 2013 11:59:30 AM UTC+1, Arkalen wrote:
> On 17/06/13 21:51, someone wrote:
> > On Monday, June 17, 2013 1:44:15 PM UTC+1, Arkalen wrote:
> >> On 17/06/13 12:38, someone wrote:

[snip]

The question was about whether the amount of changes seemed strange. As
I was thinking of beneficial changes being those that end up spreading
through the population, but had forgot about genetic drift. There does
still seem to be a related issue. It seems that random evolution relies
on mutations which are beneficial yet not regressions to earlier benefits to
explain the evolution from the common ancestor of the chimpanzee and
human to the human. This assumes that you aren't expecting the evolution to
be explainable in terms of regression to previous functionality. This
then raises the question of how much of the evolutionary change involved
these type of the mutations, and what frequency do we see such changes in
modern humans. The reason to look at the modern humans is to establish the
expected frequency of such mutations, before comparing it to what happened to
see whether there was anything surprising.

So, I am making a distinction between two types of beneficial
mutation, and asking about the frequency in modern humans of the type
that isn't a regression to functionality that had evolved earlier. How many
per person per generation *of this type* are geneticists finding do you
think?

[someone]

So you think that the genetic mutations that give humans an intellectual
advantage over chimps can in one way of looking at it be regarded as a
regression to earlier functionality?

>
> In fact a lot of the differences between humans and chimpanzees involve
> neoteny, i.e. the persistence of infantile traits into adulthood. You
> could say that walking upright was a regression by our ancestors, a case
> of arrested development.
>
> Let's imagine that our larger brains were the result of a mutation in a
> regulatory gene that leads it to inactivate a gene that would usually
> downregulate genes that promote brain development in the fetal stage.
> Would this be an advancement or a regression ? A loss or a gain of
> information ? A deactivation or an activation ?

A regression I would have thought, because at some point prior to
the regulatory gene we are imagining limiting the amount of brain
development (if I have understood you correctly), the life form
would I presume have had the larger brain. Presumably the conditions
would have had to of changed to have favoured the limiting of the brain
size and then changed again for humans such that the inhibiting of
the limiting gene was favoured.

[Arkalen]

I replied elsethread regarding this whole "regression" thing (after your
replies to Bill a few hours ago), I hope we can discuss this further there.

[Arkalen]

We don't know what those mutations are exactly but I'm willing to bet
once we find out some people will find ways to spin them that way.

>>
>> In fact a lot of the differences between humans and chimpanzees involve
>> neoteny, i.e. the persistence of infantile traits into adulthood. You
>> could say that walking upright was a regression by our ancestors, a case
>> of arrested development.
>>
>> Let's imagine that our larger brains were the result of a mutation in a
>> regulatory gene that leads it to inactivate a gene that would usually
>> downregulate genes that promote brain development in the fetal stage.
>> Would this be an advancement or a regression ? A loss or a gain of
>> information ? A deactivation or an activation ?
>
> A regression I would have thought, because at some point prior to
> the regulatory gene we are imagining limiting the amount of brain
> development (if I have understood you correctly), the life form
> would I presume have had the larger brain.

Not necessarily. For one thing, I didn't say anything about when that
regulatory gene evolved in the first place; it could have existed for as
long as brains have. Or longer, if it had a role in the regulation of
nerves. Or it could be more recent. Our earliest ancestor "prior to the
regulatory gene" could be *anything* (well, anything in our direct line
of ancestry at least).

I'm also not sure why you presume that life form prior to the regulatory
gene would have a *larger* brain; more likely it wouldn't have a brain
at all, given how upstream this purported gene is. I'm talking about a
mutation in the gene, right, not its appearance.

> Presumably the conditions
> would have had to of changed to have favoured the limiting of the brain
> size and then changed again for humans such that the inhibiting of
> the limiting gene was favoured.
>

I think you're picturing a process that's still too simple. You seem to
think that if there are genes that keep the brain small, that means that
brains used to be bigger until those genes that keep the brain small
evolved, and deactivating the genes that keep the brain small means
returning to that earlier state. Is that correct ?

That's not how development works however. Think of it; you have an
embryo, that grows, its organs develop gradually, until it becomes a
free-living juvenile, continues to grow and develop until it's an adult,
and then senesces and dies.

During this process organs grow, grow faster, grow more slowly, stop
growing, grow in different ways, change function, shrink and even
disappear altogether (though that last one mostly happens during
embryological development), at different times in the organism's lifespan.

This requires genes to start growth, to promote growth, to slow growth,
and to stop growth, all coexisting and functioning in the same organism,
*whatever the size of its organs*.

In other words, if an organisms has genes that slow down brain
development, that doesn't mean its ancestors had bigger brains before it
evolved those genes. All it means is that its brain isn't an
indefinitely-growing tumor.

You'll also notice it means you could make its brain bigger, or smaller,
or a different shape, *only by deactivating genes* at key points in its
development.

Basically, I've seen people argue that changes don't count because
they're a "regression" (or a loss of information) on a macroscopic level
(it seems to be what you're saying with the German baby, though I'll
note you don't know that. We don't know that its muscles are set up the
same way any human ancestor's have ever been), and argue that changes
don't count because they're a regression or loss of information on a
genetic level (they involve the inactivation of a gene for example).

The problem is you can't do both, because those measures are independent
of one another. And if you do do both, then all those changes that
"don't count" alone could account for most of the differences we observe
between actual organisms.

[Arkalen]

I forgot, here is some more detail on the Italian mutation :
http://www.lbl.gov/Science-Articles/Archive/LSD-Milano-Bielicki.html

You haven't explained what makes you think this (or the HIV-resistance
gene Rolf brought up) would be a regression to an earlier form. (to be
fair, what you actually said that you couldn't tell that it *wasn't* a
regression, but you could say that of ANY mutation we didn't know
everything about so I'll need a bit more than that)

[drose...@yahoo.com]

No. The common ancestor would have had a smaller brain
because it had a smaller skull. The regression would have been
in that part of our "biological clock" that determines the relative
proportions (i.e., shape) of our bodies. The part of the biological
clock that determines over all size would either be static or
would advance.

For issues like this, I recommend the book,
"Ontogeny and Phylogeny" by Stephen GouldBelknapp-Harvard, 1977).
There are some other fine descriptions in comparative anatomy
text books. However, I this book presents a popularized version
of the theory.

I also like it because there is less "preaching" in
this book relative to Gould's other books. Gould has a
tendency to digress into sermons on secular humanism. There
is some of that here, but less of it.

You made two mistakes. First, you are assuming that a brain can't
grow when a growth rate is slowed down. Second, you are assuming that
information is lost when a growth rate is slowed down.

First, there is more than one growth rate. In all animals,
there are three biological clocks that are loosely synchronized
with each other. The first biological clock sets the rate at
which our total mass increases. The second sets the rate at
which the shape of our bodies advances. The third sets the
rate at which our sexual organs mature. The hypothesis that Gould
presents is that separate networks of genes control all three.

Neotony is when the second biological clock and third
clock are slowed down. The first clock does not slow down. The
growth of the human brain is large because of mutations causing
neotony.

In all vertebrates, the baby has an anterior (i.e., head) that is relatively large compared to the posterior (i.e., butt). As the body
ages, the ratio of head size to butt size decreases.

The brain size is determined by a convolution of the first clock
and the size of the skull. The brain grows at a rate given by the first
clock up to the size allowed by the skull. In an animal that grows in
weight very fast, the size of the brain grows very fast. However,
regulation stops the brain from outgrowing the skull. A brain too large
for the skull kills the organism, so natural selection doesn't allow it.

In the common ancestor of chimpanzees and humans, the brain
was mostly constrained by the size of the skull. Because the second
clock was fast, the skull didn't grow large enough to hold a large
brain. The animal grew fast, and the brain grew with it, up to the
size allowed by the skull. However, the absolute size of the brain
was small compared to the brain human's have today.

Mutations occurred which slowed down the second clock without
slowing down the first clock. Thus, the ratio between the size of
the head and the size of the butt remained large. However, the
total weight of the adult human did not go down. In fact, it even
went up a bit. So the mutation made the brain grow larger by slowing
down the shape maturing process.

Second, you are wrong by assuming that slowing down is associated
with loss of complexity. According to your definition of information,
neotony is a loss of information because it involves two clocks slowing
down. The rate of sexual maturation and the rate of changing shape have slowed down.


Please note that slowing down is not a loss of information. Genes
that slow down things are not necessarily less complex than genes that speed up things It is a matter of changing the sensitivity of the switch. One can
not tell from the complexity of the genome which growth rates are slowed
down, which speed up, and which stay the same.

The vertebrates with the most DNA per cell are actually the
amphibians. Extant amphibians on average have smaller bodies and
smaller brains than the average extant mammal. However, they have
more DNA. This may be associated with the number of discrete changes
that the amphibian goes through.

Amphibians go through metamorphosis. They change in many ways as they
pass through one stage to another. So there have to be genes that switch
on at certain stages of growth and switch off afterwards. None of these
changes have resulted in bigger brains, at least for extant amphibians.

It appears to me that you are not a mathematician. A mathematician
would insist on a rigorous definition of "information". One can not
determine the difference between regression, advancement and equilibrium
without a rigorous and unambiguous definition of "information" or
"complexity".

You apparently have read a lot of Creationist literature. The people
who write these things are generally not mathematicians or even
scientists. So these guys tend to make up "scientific laws" and "definitions"
as they write.

Some of these people are engineers. I think engineers are very smart in
general. However, there is one short coming to engineers. Most of their
work involves controlling nature. They don't pay attention to nature. Some spend their lives "designing" things without once observing nature.

So when an engineer uses words like "information" or "complexity",
he very often is thinking about some project that he did. Engineers seldom
make a project that involves looking at something that is uncontrolled.
Even a lowly physicist, like myself, occasionally has to examine a system
that is not controlled. Some engineers only look at controlled systems.

In any case, the Creation websites are not really written with a
scientific point of view, no matter how often they make that claim.
It is written by people who have this bias toward control. They think
everything should be controlled, and trust conjectures where everything
is controlled.

"Information" does not mean "control". Please remember that.

[drose...@yahoo.com]

On Tuesday, June 18, 2013 12:37:55 PM UTC-4, drose...@yahoo.com wrote:
I give up. This message looked yukkky. However, I was trying to trim
it with hard returns. I want to be on the new Google with Word Wrap.

1) Is it possible to change from an old version of Google to a new
version of Google without joining Google all over again?

2) Does anyone know how to change an old version of Google to a new
version of Google?

3) Could someone remind me how to join Google from scratch with the
new version?

Note that my shorter lines with Word Wrap won't be a regression.
I will have simply substituted soft returns for hard returns. Shorter
does not mean regression. It sometimes means more control.

[jillery]

Are you sure you're using Old GG? I ask because Old GG isn't known to
cause the problems your posts show. You can tell you're using Old GG,
because somewhere on the screen will be a box suggesting that you
"Switch to New Google Groups".

If in fact you're using New GG, as I suspect you are, you can use the
following line in any link:

http://groups.google.com/group/talk.origins/topics?&noredirect=true

The above is from a post by Friar Broccoli.

[someone]

On Tuesday, June 18, 2013 4:31:53 PM UTC+1, Arkalen wrote:
> On 18/06/13 15:46, someone wrote:
> > On Tuesday, June 18, 2013 12:45:28 PM UTC+1, Arkalen wrote:

[snip]

>
> >> Let's imagine that our larger brains were the result of a mutation in a
> >> regulatory gene that leads it to inactivate a gene that would usually
> >> downregulate genes that promote brain development in the fetal stage.
> >> Would this be an advancement or a regression ? A loss or a gain of
> >> information ? A deactivation or an activation ?
> > A regression I would have thought, because at some point prior to
> > the regulatory gene we are imagining limiting the amount of brain
> > development (if I have understood you correctly), the life form
> > would I presume have had the larger brain.
>
> Not necessarily. For one thing, I didn't say anything about when that
> regulatory gene evolved in the first place; it could have existed for as
> long as brains have. Or longer, if it had a role in the regulation of
> nerves. Or it could be more recent. Our earliest ancestor "prior to the
> regulatory gene" could be *anything* (well, anything in our direct line
> of ancestry at least).
>
> I'm also not sure why you presume that life form prior to the regulatory
> gene would have a *larger* brain; more likely it wouldn't have a brain
> at all, given how upstream this purported gene is. I'm talking about a
> mutation in the gene, right, not its appearance.
>
>

So the regulatory gene had some function other than limiting growth brain
growth before hand, and it just happened that when the organism developed
a brain that it was limiting the brains growth?

> > Presumably the conditions
> > would have had to of changed to have favoured the limiting of the brain
> > size and then changed again for humans such that the inhibiting of
> > the limiting gene was favoured.
>
> I think you're picturing a process that's still too simple. You seem to
> think that if there are genes that keep the brain small, that means that
> brains used to be bigger until those genes that keep the brain small
> evolved, and deactivating the genes that keep the brain small means
> returning to that earlier state. Is that correct ?
>

Yes that was what I thought the scenario you were describing was indicating.

>
> That's not how development works however. Think of it; you have an
> embryo, that grows, its organs develop gradually, until it becomes a
> free-living juvenile, continues to grow and develop until it's an adult,
> and then senesces and dies.
>
> During this process organs grow, grow faster, grow more slowly, stop
> growing, grow in different ways, change function, shrink and even
> disappear altogether (though that last one mostly happens during
> embryological development), at different times in the organism's lifespan.
>
> This requires genes to start growth, to promote growth, to slow growth,
> and to stop growth, all coexisting and functioning in the same organism,
> *whatever the size of its organs*.
>
> In other words, if an organisms has genes that slow down brain
> development, that doesn't mean its ancestors had bigger brains before it
> evolved those genes. All it means is that its brain isn't an
> indefinitely-growing tumor.
>

Sorry still not with you there. Are you saying that without the genes that
the brain would be an indefinitely-growing tumor?

Since it is a scenario you are making up, why don't you fill in the earlier
picture that you are imagining?

>
> You'll also notice it means you could make its brain bigger, or smaller,
> or a different shape, *only by deactivating genes* at key points in its
> development.
>

Yes but in the scenario you are mentioning the gene restricts the size of the
brain (absolutely or relative to its body size or whatever), so it seems
reasonable to assume that without it there wouldn't be the restriction.

> Basically, I've seen people argue that changes don't count because
> they're a "regression" (or a loss of information) on a macroscopic level
> (it seems to be what you're saying with the German baby, though I'll
> note you don't know that. We don't know that its muscles are set up the
> same way any human ancestor's have ever been), and argue that changes
> don't count because they're a regression or loss of information on a
> genetic level (they involve the inactivation of a gene for example).
>
> The problem is you can't do both, because those measures are independent
> of one another. And if you do do both, then all those changes that
> "don't count" alone could account for most of the differences we observe
> between actual organisms.

As I have said in an earlier post to you it seems that random evolution relies

on mutations which are beneficial yet not regressions to earlier benefits to
explain the evolution from the common ancestor of the chimpanzee and
human to the human. This assumes that you aren't expecting the evolution to
be explainable in terms of regression to previous functionality. This
then raises the question of how much of the evolutionary change involved
these type of the mutations, and what frequency do we see such changes in
modern humans. The reason to look at the modern humans is to establish the
expected frequency of such mutations, before comparing it to what happened to
see whether there was anything surprising.

So, I am making a distinction between two types of beneficial
mutation, and asking about the frequency in modern humans of the type

that isn't a regression to functionality that had evolved earlier. Maybe the
German baby wasn't a regression, but it did seem to be the result of inhibiting
a presumably previously advantageous gene, rather than forming a new protein that performs some new functionality for example. I'm still not clear on what
problem you see with making the distinction.

[drose...@yahoo.com]

I am very sorry that I swore not to clip. Now, I want to deal with your statements one
mistake at a time. However, I can�t remove all the wrong statements from the post.

Your requirement not to clip is a type of censorship. The probability that they won�t
be posted increases with the length of the post. Thus, the moderator may end this thread
before

However, I want to mention one.

Regression to a previous state of an organism is highly improbable even over
geologically long time spans. This is my restatement of Dollo�s Law.

Please not that by regression I do NOT mean simplification. By regression, I mean
going back to a previous state regardless of its level of complexity. A tape worm may be
simpler than a planaria, from which it may have descended. However, a tape worm will
never evolve into a planaria by Dollo�s Law. The same goes for a lot of parasites.
Parasites are not �regressions� of free living organisms. The ancestor of the planaria did
not look like a tape worm.

By Dollo�s Law, the regressions that you talk about are highly improbable. Scientists have looked into exceptions to Dollo�s Law. Very few have been
found, if any. For example, whales did not develop gills to live under water.
Nor is the ear-nose-throat system divided up in whales like it is in fish.



Dollo�s Law basically says that evolution can�t reverse itself. Since the
word reverse is slightly ambiguous, various scientists have tried to refine
Dollo�s Law to reduce the ambiguity. Above is my spin on Dollo�s Law. However,
the general idea should seem plausible even without refinements.


In any generation of an organism, there are millions of historical paths
leading from the specific state of an organism. However, there may be only a
few hundred historical paths that can lead to a particular genome. Therefore,
the probability of an organism going back precisely to a previous state is
astronomically small. This is basically Dawkin�-Dollo Law.

As an organism evolves, it evolves features that block the evolution
backwards to a previous state. This is basically the Gould-Dollo Law.

Here is a link to Dollo�s Law. Note that I believe in more in Richard Dawkins
version of Dollo�s Law than in Stephen Goulds version. One reason is that there
are exceptions to Gould�s version of Dollo�s Law that have been demonstrated.
I think that the Dawkins-Dollo Law is self-evident to anyone looking at the
genome sequence of an animal.


http://en.wikipedia.org/wiki/Dollo%27s_law_of_irreversibility
�Dollo's law of irreversibility (also known as Dollo's law and Dollo's
principle) is a hypothesis proposed in 1893[1] by French-born Belgian
paleontologist Louis Dollo which states that evolution is not reversible. This
hypothesis was first stated by Dollo in this way: "An organism is unable to
return, even partially, to a previous stage already realized in the ranks of
its ancestors."[2] According to this hypothesis a structure or organ that has
been lost or discarded through the process of evolution will not reappear in
exactly the same form in that line of organisms.[3][4] According to Richard
Dawkins, the law is "really just a statement about the statistical
improbability of following exactly the same evolutionary trajectory twice (or,
indeed, any particular trajectory), in either direction."[5]
Stephen Gould viewed the idea less strictly, suggesting that "irreversibility"
forecloses certain evolutionary pathways once broad forms have emerged: "[For
example], once you adopt the ordinary body plan of a reptile, hundreds of
options are forever closed, and future possibilities must unfold within the
limits of inherited design."[6]
�
Although the exact threshold for violations of Dollo's law is unclear, there
are several case studies whose results dispute the validity of some
interpretations. For example, many taxa of gastropods have reduced shells, and
some have lost coiling of their shell coiling altogether.[8] According to S. J.
Gould's interpretation of Dollo's law, it would not be possible to regain a
coiled shell after the coiling has been lost. Nevertheless, a few
genera in the slipper snail family (Calyptraeidae) may have changed their
developmental timing (heterochrony) and regained a coiled shell from a
limpet-like shell.[8][9]�

[Arkalen]

No. I'm saying that a multicellular organism that doesn't have genes to
limit the growth of *anything* isn't a multicellular organism at all,
it's a tumor. (or a unicellular colony, to be less loaded about it).

I was using hyperbolic language but the point is that genes that limit
the growth of organs exist in all organisms that have organs. Their
existence doesn't imply the organism has a particularly small version of
said organ.

The size of the organ depends as much on the timing of those genes'
activation as on their existence (or amount, or efficacy).

>
>>
>> You'll also notice it means you could make its brain bigger, or smaller,
>> or a different shape, *only by deactivating genes* at key points in its
>> development.
>>
>
> Yes but in the scenario you are mentioning the gene restricts the size of the
> brain (absolutely or relative to its body size or whatever), so it seems
> reasonable to assume that without it there wouldn't be the restriction.

Without it the organism would die in development. There is no "without
it". I'm talking about a *modification* in an *existing* regulatory gene
that has various activating and deactivating effects on genes downstream
from it. Those genes are also already-existing. They exist in all
organisms because development involves starting and stopping various
processes at certain points in time, hence the existence of both genes
that promote the process and genes that repress it.

I'm confused as to what's in your mind when you talk about "there
wouldn't be the restriction" on the size of the brain. Either a brain
grows indefinitely (our tumor scenario), or its size is restricted. No
brain grows indefinitely. All brains' sizes are restricted. I can't
figure out whether you're actually thinking of an indefinitely-growing
brain when you consider how the organism would be without the gene that
"restricts the size of the brain" (note I did not mention a gene that
did that or did it alone), or whether you think the brain has some
proper size it ought to be and the gene is making it smaller than that.

>
>
>> Basically, I've seen people argue that changes don't count because
>> they're a "regression" (or a loss of information) on a macroscopic level
>> (it seems to be what you're saying with the German baby, though I'll
>> note you don't know that. We don't know that its muscles are set up the
>> same way any human ancestor's have ever been), and argue that changes
>> don't count because they're a regression or loss of information on a
>> genetic level (they involve the inactivation of a gene for example).
>>
>> The problem is you can't do both, because those measures are independent
>> of one another. And if you do do both, then all those changes that
>> "don't count" alone could account for most of the differences we observe
>> between actual organisms.
>
> As I have said in an earlier post to you it seems that random evolution relies
> on mutations which are beneficial yet not regressions to earlier benefits to
> explain the evolution from the common ancestor of the chimpanzee and
> human to the human.

Evolution doesn't care whether a mutation is a "regression to earlier
benefits" or not. In fact there are many cases of organisms "regressing
to earlier benefits". It happens (though they never "regress" to exactly
what the "earlier benefits" were)

You're right that it's not the only thing that happens, but to actually
show this we need to define what "regression to earlier benefits" *means*.

> This assumes that you aren't expecting the evolution to
> be explainable in terms of regression to previous functionality.

Most of it is actually... if you define "regression to previous
functionality" as widely as I've seen some creationists do.

> This
> then raises the question of how much of the evolutionary change involved
> these type of the mutations, and what frequency do we see such changes in
> modern humans. The reason to look at the modern humans is to establish the
> expected frequency of such mutations, before comparing it to what happened to
> see whether there was anything surprising.
>
> So, I am making a distinction between two types of beneficial
> mutation, and asking about the frequency in modern humans of the type
> that isn't a regression to functionality that had evolved earlier. Maybe the
> German baby wasn't a regression, but it did seem to be the result of inhibiting
> a presumably previously advantageous gene, rather than forming a new protein that performs some new functionality for example. I'm still not clear on what
> problem you see with making the distinction.
>

There are a number of problems with that distinction.
One problem is that new genes forming new proteins that perform some new
functionality are a tiiiiiiiiiiiiiny proportion of the actual
significant changes that happen in multicellular organisms. Changes in
regulatory genes are much more important. Deactivation of genes are also
very important. Also, all "new genes" are modifications of a
previously-existing gene. I don't think you can get a new, unique gene
in a single step. You can get one in two steps - duplication of a gene,
modification of the duplicate, and all the variations on that theme. But
the point is if you say "new protein" excludes "modification of an old
protein" (which some do, I don't know if you're one of them) you're
again ignoring most of actual evolution that happens. In fact if you go
from there to excluding "genetic duplication" and "modification of an
inactive strech of genome" you've gone and excluded all evolutionary
processes. Congratulations. Now you're stuck with the fact all of the
processes you've excluded as "not what's needed for random evolution"
can fully account for the history of life post-abiogenesis.

The other problem with excluding 99% of the genetic mechanisms of most
actual evolutionary events is that the remainder will necessarily be
rare. Which is convenient for you, but makes it a lot less significant
when we find few examples.

Speaking of, I don't know if you've read my next post yet but is
Apolipoprotein A-1 Milano a "new protein" with "new functionality" ? Why
or why not ?

Finally, attn biologists : I've tried to look up the differences between
the chimpanzee proteome and the human one but couldn't find much. What
proteins do we have that chimpanzees don't or vice-versa, really ?

[someone]

On Tuesday, June 18, 2013 8:59:41 PM UTC+1, Darwin123 wrote:
> On Tuesday, June 18, 2013 3:16:58 PM UTC-4, someone wrote:
>

[snip]

Two things, the first is that I don't mind people snipping, I snip myself, it
is sensible to prevent posts being too long. So the accusations of censorship
etc. seem overly dramatic. I'm just ask people to write [snip] when they do.
This prevents the posts being effectively re-written, with points being removed
etc., without it be indicated to the reader that the text has been edited
somewhat. The second is that no one is talking about going back to *precisely*
an earlier genetic state. For example one mutation might have occurred which
allowed what was some junk DNA to produce a protein of some kind. Another
mutation might then occur which breaks that protein production, but the second
needn't put it back to the *precise* state it was before the advantageous
mutation, but the DNA section in this example is returned back to the same
*functionality* that it had before (not producing the protein), not necessarily
the precisely the same base pair sequence.


[snip]

[someone]

> actual evolutionary events is that the remainder will necessarily be
> rare. Which is convenient for you, but makes it a lot less significant
> when we find few examples.
>
> Speaking of, I don't know if you've read my next post yet but is
> Apolipoprotein A-1 Milano a "new protein" with "new functionality" ? Why
> or why not ?
>
> Finally, attn biologists : I've tried to look up the differences between
> the chimpanzee proteome and the human one but couldn't find much. What
> proteins do we have that chimpanzees don't or vice-versa, really ?

Ok I'll just address the points in one place as you seem to be having trouble
following what I'm saying. I'm saying that if a mutation simply inhibits
earlier functionality then while there might be environments where it could be
beneficial the benefit would be by a recessive mutation. Now let's take your
brain regulation example. Let us suppose that the brain size was a certain
proportion of the body size, and that a mutation occurred that made it
a lower proportion of the overall size, but that in the environment in question
this was beneficial, but then at a later date, the environment had changed
and the mutation that had made it a lower proportion of overall size got
corrupted, then we could consider this as a functional regression as far
as these genes are concerned back to earlier functionality. If there was
another example where certain timings of when genes were turned on and turned
off were significant and a mutation occurred which changed a production
sequence to one that had never was in the history of the organism, and that
was beneficial, then I would consider this a new beneficial mutation which
wasn't a regression.

Now you might question what about a change where the organism had changed
through mutation and the corruption of a gene exposes some functionality
of these changes that the gene had previously covered up, so that the
functionality that is exposed is new, but it was done through the corruption
of a previously beneficial gene. In such cases I would consider the genes
that encoded the new functionality to have encoded new beneficial mutations,
but that the corruption of the gene that was inhibiting such functionality
a regressive mutation even though it didn't cause on the larger scale a
return to existing functionality. So new regulatory genes, or new improvements
to regulation => new beneficial mutations. Whereas previous improvements removed
by reproduction error and the *gene* returning to functionality it previously
had in the life would be a regressive mutation, even if it turned out to be
beneficial. So I'm not limiting new beneficial mutations to those that
produce new proteins.

As for the Italian example, I don't know, as I said it wasn't clear from the
information on it, though the German example did seem like it could be one
of the more frequent regressive mutations.

So how much of the human genome were you thinking was displaying what
could be considered new beneficial mutations as I am rather crudely describing
them?

[someone]

Just thought I'd say is that it should be "by a regressive mutation", with
what is meant by regressive being discussed.

[Arkalen]

I am *definitely* having trouble following what you are saying... What
do recessive mutations have to do with this ? Are you confusing
"recessive" with "regression" ?

And you still have not understood that a mutation inhibiting something
doesn't mean it's restoring an earlier functionality. The whole body
functions though genes inhibiting and promoting each other; adding an
inhibition to that will yield a *new* result, not return to an earlier one.

> Now let's take your
> brain regulation example. Let us suppose that the brain size was a certain
> proportion of the body size, and that a mutation occurred that made it
> a lower proportion of the overall size, but that in the environment in question
> this was beneficial, but then at a later date, the environment had changed
> and the mutation that had made it a lower proportion of overall size got
> corrupted, then we could consider this as a functional regression as far
> as these genes are concerned back to earlier functionality.

Yes. But that is not the situation my brain regulation example
describes. I have not said anything about what size or proportion the
brain was at any point earlier in the lineage, that's something you're
adding to the scenario.

If you insist I can specifically posit that at the moment the mutation
happened the organism's brain was as large (or its encephalization
quotient was as large) as it had ever been for any of that organism's
ancestors. It wouldn't make the example any less realistic given that's
probably what actually happened.

> If there was
> another example where certain timings of when genes were turned on and turned
> off were significant and a mutation occurred which changed a production
> sequence to one that had never was in the history of the organism, and that
> was beneficial, then I would consider this a new beneficial mutation which
> wasn't a regression.

Okay.

>
> Now you might question what about a change where the organism had changed
> through mutation and the corruption of a gene exposes some functionality
> of these changes that the gene had previously covered up, so that the
> functionality that is exposed is new, but it was done through the corruption
> of a previously beneficial gene. In such cases I would consider the genes
> that encoded the new functionality to have encoded new beneficial mutations,
> but that the corruption of the gene that was inhibiting such functionality
> a regressive mutation even though it didn't cause on the larger scale a
> return to existing functionality.

That doesn't make sense to me. What does "regressive mutation" mean if
it doesn't cause a return to existing functionality ? It's not as if it
causes a small-scale return to existing functionality either; there is
no reason to think the corrupted gene functions the same way any of its
ancestors did. Some genes arise de novo from a mutation in
otherwise-useless DNA, but most arise from the modification of
previously-existing genes, so making them nonfunctional would be in no
way a reversion to previous (non)functionality.

> So new regulatory genes, or new improvements
> to regulation => new beneficial mutations. Whereas previous improvements removed
> by reproduction error and the *gene* returning to functionality it previously
> had in the life would be a regressive mutation, even if it turned out to be
> beneficial. So I'm not limiting new beneficial mutations to those that
> produce new proteins.

While there are still some things I'm confused about (see higher up) I
think I mostly understand your definition of "new beneficial mutations"
and I think it's a definition we can work with.

>
> As for the Italian example, I don't know, as I said it wasn't clear from the
> information on it, though the German example did seem like it could be one
> of the more frequent regressive mutations.

What further information would you need on the Italian example before
you could say anything about it ? Maybe it's something I could find.

>
> So how much of the human genome were you thinking was displaying what
> could be considered new beneficial mutations as I am rather crudely describing
> them?
>

Most of the beneficial mutations in the human genomes vs our common
ancestor with chimpanzees, whatever proportion of the total mutations
that is, would be "new" by your definition, given obviously none of our
ancestors anywhere in our lineage differed from other apes the way we
do. We're the first bipeds in our lineage, the first hairless mammals,
have the highest encephalization quotient in a lineage that's shown
steady increase in encephalization quotient... Possibly the loss of our
muscles and other adaptations to an arboreal lifestyle could be
considered "regressions" by your definition. Even though the loss of our
jaw muscles in particular can't be separated from the development of our
brains and speech, so even what you'd call regressions are important.

Hey, is it me or is the Wikipedia page on the chimpanzee genome project
more detailed than it was last time I looked ? According to it they
identified 600 genes that underwent strong selection in humans and
chimpanzees; we can call those the "beneficial mutations". They don't
give numbers on which genes do what, but nearly all those they mention
seem like they'd be non-regressions by your definition.

[Darwin123]

On Wednesday, June 19, 2013 3:25:51 AM UTC-4, someone wrote:
-
-clip-
-

> So how much of the human genome were you thinking was displaying what
> could be considered new beneficial mutations as I am rather crudely
>describing them?

I just read an article in Discover that claimed there are on average
100-104 point substitution mutations per human being per generation. The
articles that you read seem to imply that there are about 3-4 mutations
per human being per generation that eventually spread through the entire
population.

I assume that the mutations that spread through the population were
slightly beneficial either to the gene itself or the individual. If they
weren't beneficial, their spread would violate the Hardy-Weinberg law. I
assume that the remaining mutations were either neutral or harmful to either
the gene or the individual.

Therefore, I tentatively speculate that the percentage of point
substitutions that are beneficial in some way is about 3%. The article
conjectures that we may be evolving new capabilities, but does not
conjecture on what these capabilities may be.

There are some huge qualifications to this. When I say it spread
through the population, I am assuming that it took a very long time to
spread through the population. When I say beneficial, I mean very slightly
beneficial.

As to whether a change was a regression or not, I don't know. I do not
know what you mean by a regression. A rapid Google search indicates that
it is a word sometimes used by Creationists with reference to some type of
mutation. However, their use of the word is unclear. My bias is that the word
"regression" is being used an arbitrary assignment of value.


The word "regression" is not part of the scientific jargon.
I have never read the word "regression" used in a scientific article.
I have seen the word "atavism" used in scientific articles. Maybe
regression is being used to refer to some type of atavism. However,
"atavism" is not applied to the gene itself. It refers to the phenotype
expressed by the gene. There is no way to tell by looking at a gene
sequence whether or not it is an atavism.

A creature can have a large skull that is very thick. Just because
the skull may have been relatively bigger in days of yore doesn't mean
it had a relatively heavier brain. Furthermore, the brain in ancient
days may not have been as convoluted as the brains of extant humans.
The brains of our ancestors may have also included more white matter
than grey matter. Hence, an atavism that would increase the relative
head size, if such existed, need not have returned any "function" to
the brain.

The benefit or harm can only be determined in context of
the environment. Maybe in the Miocene, a thick skull was more
beneficial than having a large brain. The word "benefit" is only
defined in context of the environment.

My previous statement has presuppositions as to what you mean by
"function", of course. The word "function" is used in scientific literature.
However, it is usually attached to a long essay on what it means to the
writer. I just finished two books by Konrad Lorenz. He spends some time
defining "function", with a little discussion on how other scientists
define it.

Lorenz defines "function" in terms of natural selection. A function
to him is a phenotype together with a state in the environment that favors
the survival and reproduction of the organism. He makes special emphasis
that one can not define a function separately from the immediate
environment that the organism lives in. It seems to me that your definition
of "function" is totally independent of the environment that the organism
lives in.

The same goes for the words "beneficial" and "harmful". They have
meaning only in the context of the environment that the organism lives
in. Environments are very complex. Hence, the only falsifiable way to
determine "benefit" is by determining whether the gene spreads through the
population. This is going to vary with the environment. There is no way
to look at the sequence of a gene and determine its benefit.

You use the word "function" rather glibly, and expect everyone else
to automatically know what you are talking about. Again, a rapid Google
search indicates that the words "function" and "functionality" are often
used by Creationists. Again, this causes my prejudice memes to express
themselves. It appears to me that the the concept of "functionality"
is being used rather arbitrarily.

Scientists can determine two things about a gene sequence without
knowing the environment of the organism. Scientists can determine what
the gene sequence is, and whether at any given time it is expressed.
Their are now special tests for determining whether a specific gene
is "on" or "off". These tests do not depend on the phenotypic result
of this gene. So really, all we can talk about without knowing the
environment of the organism are the gene sequence and its state of
expression.

If you want to tell us what you mean by "function", then please tell
us in terms of gene sequence, gene expression and the environment the
organism lives in. You can make very general statements about all three,
but you have to include all three.

[John Harshman]

On 6/19/13 7:03 AM, Darwin123 wrote:
> On Wednesday, June 19, 2013 3:25:51 AM UTC-4, someone wrote:
> -
> -clip-
> -
>> So how much of the human genome were you thinking was displaying what
>> could be considered new beneficial mutations as I am rather crudely
>> describing them?
>
> I just read an article in Discover that claimed there are on average
> 100-104 point substitution mutations per human being per generation. The
> articles that you read seem to imply that there are about 3-4 mutations
> per human being per generation that eventually spread through the entire
> population.

I doubt that last figure sincerely. I suspect the average is very near
zero. In fact neutral theory predicts that the number of mutations fixed
per generation is equal to the mutation rate, i.e. around 100 total per
generation (given a stable population, which we don't have). So the
number per person is about 100 divided by the population size.

> I assume that the mutations that spread through the population were
> slightly beneficial either to the gene itself or the individual. If they
> weren't beneficial, their spread would violate the Hardy-Weinberg law. I
> assume that the remaining mutations were either neutral or harmful to either
> the gene or the individual.

You misunderstand Hardy-Weinberg. One assumption of Hardy-Weinberg is
that the population is infinite. Do I have to explain that this
assumption doesn't hold? In any real population, frequencies fluctuate
stochastically, and neutral mutations must inevitably become either
fixed or extinct, given enough time.

> Therefore, I tentatively speculate that the percentage of point
> substitutions that are beneficial in some way is about 3%. The article
> conjectures that we may be evolving new capabilities, but does not
> conjecture on what these capabilities may be.

I'm afraid your "therefore" is a result of false premises, and I've
touched on just a few of them above.

[someone]

On Wednesday, 19 June 2013 13:08:31 UTC+1, Arkalen wrote:
> On 19/06/13 08:25, someone wrote:
> > On Tuesday, June 18, 2013 9:07:59 PM UTC+1, Arkalen wrote:
> >> On 18/06/13 20:16, someone wrote:
> >>> On Tuesday, June 18, 2013 4:31:53 PM UTC+1, Arkalen wrote:
> >>>> On 18/06/13 15:46, someone wrote:
> >>>>> On Tuesday, June 18, 2013 12:45:28 PM UTC+1, Arkalen wrote:
> >>> [snip]

[snip]

> > Ok I'll just address the points in one place as you seem to be having trouble
> > following what I'm saying. I'm saying that if a mutation simply inhibits
> > earlier functionality then while there might be environments where it could be
> > beneficial the benefit would be by a recessive mutation.
>
> I am *definitely* having trouble following what you are saying... What
> do recessive mutations have to do with this ? Are you confusing
> "recessive" with "regression" ?
>
>

No I'm not. I did make another post mentioning that it should have said
regressive.

>
> And you still have not understood that a mutation inhibiting something
> doesn't mean it's restoring an earlier functionality. The whole body
> functions though genes inhibiting and promoting each other; adding an
> inhibition to that will yield a *new* result, not return to an earlier one.
>

I didn't say it did, in fact the post went on to make it clear I wasn't
saying that.

>
>
> > Now let's take your
> > brain regulation example. Let us suppose that the brain size was a certain
> > proportion of the body size, and that a mutation occurred that made it
> > a lower proportion of the overall size, but that in the environment in question
> > this was beneficial, but then at a later date, the environment had changed
> > and the mutation that had made it a lower proportion of overall size got
> > corrupted, then we could consider this as a functional regression as far
> > as these genes are concerned back to earlier functionality.
>
>
>
> Yes. But that is not the situation my brain regulation example
> describes. I have not said anything about what size or proportion the
> brain was at any point earlier in the lineage, that's something you're
> adding to the scenario.
>

Yes, I was adding what I said following "Let us suppose.."

>
> If you insist I can specifically posit that at the moment the mutation
> happened the organism's brain was as large (or its encephalization
> quotient was as large) as it had ever been for any of that organism's
> ancestors. It wouldn't make the example any less realistic given that's
> probably what actually happened.

There's no need, what I added illustrated the point of what I meant by
a regressive mutation, though from what's written I'm not sure
that you got it, though maybe from what you said later, you did.

>
> > If there was
> > another example where certain timings of when genes were turned on and turned
> > off were significant and a mutation occurred which changed a production
> > sequence to one that had never was in the history of the organism, and that
> > was beneficial, then I would consider this a new beneficial mutation which
> > wasn't a regression.
>
>
> Okay.
>
> > Now you might question what about a change where the organism had changed
> > through mutation and the corruption of a gene exposes some functionality
> > of these changes that the gene had previously covered up, so that the
> > functionality that is exposed is new, but it was done through the corruption
> > of a previously beneficial gene. In such cases I would consider the genes
> > that encoded the new functionality to have encoded new beneficial mutations,
> > but that the corruption of the gene that was inhibiting such functionality
> > a regressive mutation even though it didn't cause on the larger scale a
> > return to existing functionality.
>
> That doesn't make sense to me. What does "regressive mutation" mean if
> it doesn't cause a return to existing functionality ? It's not as if it
> causes a small-scale return to existing functionality either; there is
> no reason to think the corrupted gene functions the same way any of its
> ancestors did. Some genes arise de novo from a mutation in
> otherwise-useless DNA, but most arise from the modification of
> previously-existing genes, so making them nonfunctional would be in no
> way a reversion to previous (non)functionality.
>

Well it could cause a small-scale return to the same functionality though
couldn't it, and not in a way that would later could be seen as a necessary
stepping stone which was new, and necessary for a later beneficial mutation.
Just for clarity, if the mutation was necessary for a later beneficial
mutation, then I would count it as a new beneficial mutation (the benefit was
that it brought the organism a step closer to the later beneficial mutation).
It just wouldn't spread like a beneficial mutation until the final steps were
made, and its benefits became manifest. Though I would expect beneficial
mutations such as these which only appear beneficial with hindsight to be
quite rare as I would expect they would generally be more likely to be
corrupted than the mutation which would have made them beneficial
occur.

>
>
> > So new regulatory genes, or new improvements
> > to regulation => new beneficial mutations. Whereas previous improvements removed
> > by reproduction error and the *gene* returning to functionality it previously
> > had in the life would be a regressive mutation, even if it turned out to be
> > beneficial. So I'm not limiting new beneficial mutations to those that
> > produce new proteins.
>
> While there are still some things I'm confused about (see higher up) I
> think I mostly understand your definition of "new beneficial mutations"
> and I think it's a definition we can work with.
>
> > As for the Italian example, I don't know, as I said it wasn't clear from the
> > information on it, though the German example did seem like it could be one
> > of the more frequent regressive mutations.
>
> What further information would you need on the Italian example before
> you could say anything about it ? Maybe it's something I could find.
>
>

What functionality did the section of DNA have before it went onto produce
the gene which then went on to be mutated. If it was the same functionality
it gets with the mutation then it is a regression.

So any information that would help answer that question might help.

>
> > So how much of the human genome were you thinking was displaying what
> > could be considered new beneficial mutations as I am rather crudely describing
> > them?
>
>
> Most of the beneficial mutations in the human genomes vs our common
> ancestor with chimpanzees, whatever proportion of the total mutations
> that is, would be "new" by your definition, given obviously none of our
> ancestors anywhere in our lineage differed from other apes the way we
> do. We're the first bipeds in our lineage, the first hairless mammals,
> have the highest encephalization quotient in a lineage that's shown
> steady increase in encephalization quotient... Possibly the loss of our
> muscles and other adaptations to an arboreal lifestyle could be
> considered "regressions" by your definition. Even though the loss of our
> jaw muscles in particular can't be separated from the development of our
> brains and speech, so even what you'd call regressions are important.
>
> Hey, is it me or is the Wikipedia page on the chimpanzee genome project
> more detailed than it was last time I looked ? According to it they
> identified 600 genes that underwent strong selection in humans and
> chimpanzees; we can call those the "beneficial mutations". They don't
> give numbers on which genes do what, but nearly all those they mention
> seem like they'd be non-regressions by your definition.

It may of changed I'm not sure. So you were saying that most of the human
genome is made up of new beneficial mutations? Because the speech ones seemed
beneficial. This seems to open up questions about what kind of % of the
differences should we expect to be of the new mutation type, and what
kind of rate per person per generation would they be estimated to be
occurring at, and what kind of rate are we observing them at now?

[Arkalen]

On 19/06/13 19:07, someone wrote:
> On Wednesday, 19 June 2013 13:08:31 UTC+1, Arkalen wrote:
>> On 19/06/13 08:25, someone wrote:
>>> On Tuesday, June 18, 2013 9:07:59 PM UTC+1, Arkalen wrote:
>>>> On 18/06/13 20:16, someone wrote:
>>>>> On Tuesday, June 18, 2013 4:31:53 PM UTC+1, Arkalen wrote:
>>>>>> On 18/06/13 15:46, someone wrote:
>>>>>>> On Tuesday, June 18, 2013 12:45:28 PM UTC+1, Arkalen wrote:
>>>>> [snip]
> [snip]
>>> Ok I'll just address the points in one place as you seem to be having trouble
>>> following what I'm saying. I'm saying that if a mutation simply inhibits
>>> earlier functionality then while there might be environments where it could be
>>> beneficial the benefit would be by a recessive mutation.
>>
>> I am *definitely* having trouble following what you are saying... What
>> do recessive mutations have to do with this ? Are you confusing
>> "recessive" with "regression" ?
>>
>>
>
> No I'm not. I did make another post mentioning that it should have said
> regressive.

Yes, I stupidly forgot to read it before clicking "post", sorry about that.

But you would agree that if it did not cause a small-scale return to the
same functionality then you wouldn't call it a regressive mutation ?

> Just for clarity, if the mutation was necessary for a later beneficial
> mutation, then I would count it as a new beneficial mutation (the benefit was
> that it brought the organism a step closer to the later beneficial mutation).

Paradoxically I think that's being *too* permissive of you. One of the
basic points of evolution is that it doesn't look ahead; it consists of
processes working in the present, not guessing at the future. Whether a
mutation caused another later mutation to be "new/beneficial" shouldn't
affect whether the first mutation was "regressive" or not.

Otherwise couldn't I argue that no mutation is regressive because a
later beneficial mutation could always come on top of them ? And the
idea that a mutation would be "regressive" and ten generation later not
be "regressive" anymore because of other mutations is rather strange.

> It just wouldn't spread like a beneficial mutation until the final steps were
> made, and its benefits became manifest. Though I would expect beneficial
> mutations such as these which only appear beneficial with hindsight to be
> quite rare as I would expect they would generally be more likely to be
> corrupted than the mutation which would have made them beneficial
> occur.

A mutation that isn't beneficial on its own but is beneficial in
combination with other mutations is basically a neutral mutation when
it's on its own, and as we discussed neutral mutations spread just fine.
The relevant probabilities isn't "what are the odds of THIS neutral
mutation spreading and then THAT beneficial mutation appearing", it's
"what are the odds that one of the neutral mutations that can be
expected to get fixed would have beneficial effects if combined with
another mutation".

>
>>
>>
>>> So new regulatory genes, or new improvements
>>> to regulation => new beneficial mutations. Whereas previous improvements removed
>>> by reproduction error and the *gene* returning to functionality it previously
>>> had in the life would be a regressive mutation, even if it turned out to be
>>> beneficial. So I'm not limiting new beneficial mutations to those that
>>> produce new proteins.
>>
>> While there are still some things I'm confused about (see higher up) I
>> think I mostly understand your definition of "new beneficial mutations"
>> and I think it's a definition we can work with.
>>
>>> As for the Italian example, I don't know, as I said it wasn't clear from the
>>> information on it, though the German example did seem like it could be one
>>> of the more frequent regressive mutations.
>>
>> What further information would you need on the Italian example before
>> you could say anything about it ? Maybe it's something I could find.
>>
>>
>
> What functionality did the section of DNA have before it went onto produce
> the gene which then went on to be mutated. If it was the same functionality
> it gets with the mutation then it is a regression.
>
> So any information that would help answer that question might help.

From what I've been able to find out, the gene involved is
Apolipoprotein A-1 (the mutated version being called Apolipoprotein A-1
Milano), which belongs to a family of genes called (surprise)
apolipoproteins. Apolipoprotein A-1 has been related to Apolipoprotein
A-II, A-IV, E, CI, CII and CIII, all of which are produced in the liver
and have roles binding lipids. They all consist of repeating
22-amino-acid sequences (themselves made to two 11-amino-acid
sequences). I'm just going to quote an article describing the
evolutionary sequence described in another paper I don't have access to:

"All the soluble apolipoprotein genes have an almost identical structure
and contain very similar internal repeats of 11 and 22 codons. Based on
this information, Luo et al. proposed an evolutionary scenario that
postulates the following: The common ancestor of the apolipoprotein
genes was very similar to the present day apoC-I in structure and
length. (...) This gene was duplicated into two; one of them led to
apoC-I and the other became the common ancestor of all of the other
apolipoprotein genes. (...) In [that] lineage, the first 11 codons and
exon 4 were duplicated and then the whole gene was duplicated into two.
One of the two genes became the present-day apoC-II, and the other, the
common ancestor of the apolipoprotein genes other than apoC-I and
apoC-II. In the latter lineage, a duplication of the first 22 codons of
exon 4 occurred and then a duplication of the whole gene followed. In
one of the two resultant genes, a deletion of the first 11 codons of
exon 4 occurred and the gene was later duplicated, one part leading to
apoA-II and the other to apoC-III. The other gene experiences six or
seven duplications of the 22 codon repeat and a duplication or deletion
of 11 codons in exon 4 and was then duplicated: one of them gained two
duplications of 22 codons and became the present day apoE; the other
gained one duplication of 22 codons in exon 4 and was then duplicated,
one leading to apoA-I and the other to apoA-IV after gaining three
duplications of 22 codons in exon 4."

(L. Chan, 1989, "The apolipoprotein multigene family: Structure,
expression, evolution, and molecular genetics", Klinische Wochenschrift)

So basically the apoA-I gene has been involved in lipid metabolism for
as long as it's existed. The Milano version differs from the human
wild-type by a single amino acid. From this article I linked to earlier :
http://www.lbl.gov/Science-Articles/Archive/LSD-Milano-Bielicki.html

It seems that apoA-I is responsible for making HDL (high-density
lipoproteins, or "good cholesterol"). The mutation means that some of
the proteins it makes will stick together (which "restricts HDL growth",
hence why people with that mutation have an HDL deficiency), but the
others will have a free bond that captures free radicals, thus
preventing the inflammation that causes heart disease.

Now, for one thing point mutations are highly unlikely to be reversed,
not on the short term at least (consider : there is a very low rate of
mutation per nucleotide; given there are a lot of nucleotides in the
genome that still sums up to a lot of mutations overall. But if you're
looking at the odds of a *specific* mutation happening at a *specific*
place (i.e. what you need for a specific mutation to be reversed) then
the probability is... that very low rate of mutation per nucleotide.
Divided by four, because we're looking for a specific base-pair in this
case), so it's extremely improbable that apoA-I Milano is a reversion to
a specific earlier form. Not only that, it's not like we have ancestors
for whom heart disease due to fatty diets was a *worse* problem than it
is for us modern humans, that they'd evolve that variant and
subsequently lose it.

>
>>
>>> So how much of the human genome were you thinking was displaying what
>>> could be considered new beneficial mutations as I am rather crudely describing
>>> them?
>>
>>
>> Most of the beneficial mutations in the human genomes vs our common
>> ancestor with chimpanzees, whatever proportion of the total mutations
>> that is, would be "new" by your definition, given obviously none of our
>> ancestors anywhere in our lineage differed from other apes the way we
>> do. We're the first bipeds in our lineage, the first hairless mammals,
>> have the highest encephalization quotient in a lineage that's shown
>> steady increase in encephalization quotient... Possibly the loss of our
>> muscles and other adaptations to an arboreal lifestyle could be
>> considered "regressions" by your definition. Even though the loss of our
>> jaw muscles in particular can't be separated from the development of our
>> brains and speech, so even what you'd call regressions are important.
>>
>> Hey, is it me or is the Wikipedia page on the chimpanzee genome project
>> more detailed than it was last time I looked ? According to it they
>> identified 600 genes that underwent strong selection in humans and
>> chimpanzees; we can call those the "beneficial mutations". They don't
>> give numbers on which genes do what, but nearly all those they mention
>> seem like they'd be non-regressions by your definition.
>
> It may of changed I'm not sure. So you were saying that most of the human
> genome is made up of new beneficial mutations?

No. 600 genes isn't a lot. It isn't a lot compared to the total number
of human genes, and it's tiny compared to the size of the human genome
overall. Remember your 20 million mutations over 5 million years between
our genome and chimpanzees' is based on all the nucleotides, not just
those that are genes AND are subject to strong selection.

> Because the speech ones seemed
> beneficial. This seems to open up questions about what kind of % of the
> differences should we expect to be of the new mutation type, and what
> kind of rate per person per generation would they be estimated to be
> occurring at, and what kind of rate are we observing them at now?
>

If all those 600 genes were new and beneficial, 600 over 5 million years
sounds hardly even detectable on a centennial timescale. There are
probably more in practice; I don't know whether the 600 was "that's how
many we've positively identified to date" or "that's how many we can
identify from looking at which parts of the genome underwent selection".

[someone]

On Wednesday, 19 June 2013 20:59:02 UTC+1, Arkalen wrote:
> On 19/06/13 19:07, someone wrote:
> > On Wednesday, 19 June 2013 13:08:31 UTC+1, Arkalen wrote:
> >> On 19/06/13 08:25, someone wrote:
> >>> On Tuesday, June 18, 2013 9:07:59 PM UTC+1, Arkalen wrote:
> >>>> On 18/06/13 20:16, someone wrote:
> >>>>> On Tuesday, June 18, 2013 4:31:53 PM UTC+1, Arkalen wrote:
> >>>>>> On 18/06/13 15:46, someone wrote:
> >>>>>>> On Tuesday, June 18, 2013 12:45:28 PM UTC+1, Arkalen wrote:
> >>>>> [snip]
> > [snip]
>

[snip]

>
> >>> Now you might question what about a change where the organism had changed
> >>> through mutation and the corruption of a gene exposes some functionality
> >>> of these changes that the gene had previously covered up, so that the
> >>> functionality that is exposed is new, but it was done through the corruption
> >>> of a previously beneficial gene. In such cases I would consider the genes
> >>> that encoded the new functionality to have encoded new beneficial mutations,
> >>> but that the corruption of the gene that was inhibiting such functionality
> >>> a regressive mutation even though it didn't cause on the larger scale a
> >>> return to existing functionality.
> >> That doesn't make sense to me. What does "regressive mutation" mean if
> >> it doesn't cause a return to existing functionality ? It's not as if it
> >> causes a small-scale return to existing functionality either; there is
> >> no reason to think the corrupted gene functions the same way any of its
> >> ancestors did. Some genes arise de novo from a mutation in
> >> otherwise-useless DNA, but most arise from the modification of
> >> previously-existing genes, so making them nonfunctional would be in no
> >> way a reversion to previous (non)functionality.
>
> > Well it could cause a small-scale return to the same functionality though
> > couldn't it, and not in a way that would later could be seen as a necessary
> > stepping stone which was new, and necessary for a later beneficial mutation.
>
> But you would agree that if it did not cause a small-scale return to the
> same functionality then you wouldn't call it a regressive mutation ?
>

Yes (though I am not saying it has to return to precisely the same
state to still have the same functionality).

>
> > Just for clarity, if the mutation was necessary for a later beneficial
> > mutation, then I would count it as a new beneficial mutation (the benefit was
> > that it brought the organism a step closer to the later beneficial mutation).
>
> Paradoxically I think that's being *too* permissive of you. One of the
> basic points of evolution is that it doesn't look ahead; it consists of
> processes working in the present, not guessing at the future. Whether a
> mutation caused another later mutation to be "new/beneficial" shouldn't
> affect whether the first mutation was "regressive" or not.
>

Well I'd consider it a new beneficial mutation for the reason I gave.

> Otherwise couldn't I argue that no mutation is regressive because a
> later beneficial mutation could always come on top of them ? And the
> idea that a mutation would be "regressive" and ten generation later not
> be "regressive" anymore because of other mutations is rather strange.
>

But the step would have to be one in which was necessary for the later
mutation. So that in hindsight it would have to be counted as a necessary
mutation for the benefit. It would belong to the set of mutations required
for the benefit. I can see your point about the potential for people to claim
that all the regressive mutations are ones that will be in a beneficial mutation
set in the lifeform's future evolution, but we could presumably examine how many necessary steps which provide no functional benefit are on average
associated with a new beneficial mutation which is showing its benefit. If
the issue focuses in on whether a particular current mutation is regressive
or not, then I would have thought that a classification of "presumed regressive"
would be fine. If later it turned out to be a necessary step to a new mutation
its classification could be changed. I don't think it would be that surprising
that it was presumed regressive, or that it would be strange if it turned out not to be, that it wasn't originally classified as a beneficial mutation.

> > It just wouldn't spread like a beneficial mutation until the final steps were
> > made, and its benefits became manifest. Though I would expect beneficial
> > mutations such as these which only appear beneficial with hindsight to be
> > quite rare as I would expect they would generally be more likely to be
> > corrupted than the mutation which would have made them beneficial
> > occur.
>
> A mutation that isn't beneficial on its own but is beneficial in
> combination with other mutations is basically a neutral mutation when
> it's on its own, and as we discussed neutral mutations spread just fine.
> The relevant probabilities isn't "what are the odds of THIS neutral
> mutation spreading and then THAT beneficial mutation appearing", it's
> "what are the odds that one of the neutral mutations that can be
> expected to get fixed would have beneficial effects if combined with
> another mutation".
>

You're right, I had overlooked that there could be more than one candidate
that could potentially be mutated into something that combined with that
particular mutational change that had seemed neutral to provide the
lifeform with an advantage in natural selection, rather than just mess
the bio-mechanism. I was thinking that it would seem quite remarkable,
and that things would be lot more likely to just mess up, so chances of
something just being mutated into something that would react directly
or indirectly with it or its product wouldn't be enough.

Also I now realise that not only did I assume only one candidate, but
I assumed it would be more than 1 mutation away from just encoding
whatever would be needed to be encoded for something to react with what
seemed like a neutral mutation which nothing else which had been developed
seemed to have any particular reaction to, and to do so to create a new
beneficial mutation. Remove that assumption and then I wouldn't expect
the neutral mutation to be more likely corrupted. Do you know of a
reason to remove that assumption? The reason I had it was because I saw no
reason for any bias towards encoding whatever it was that would have been
beneficial reacting directly or indirectly with the neutral mutation that got fixed.

I read that it said:
---
The mutated protein, however, boasts an antioxidant in the form of a sulfur-based residue that mops up unpaired electrons and prevents them from triggering arterial inflammation
---
If a protein which boasted such an antioxidant wasn't present before in the
evolutionary history of their genome then it would seem like a new mutation.
Though whether it is beneficial or not isn't clear. The article mentions
that:
---
The trick is to develop a simple pharmaceutical peptide that targets the upregulation of the ABCA1 receptor, exports cholesterol like the conventional protein, and fights arterial wall oxidation like the Milano mutation..
---

What I'm not clear on whether not cleaning out the cholesterol of the cells
as efficiently as the conventional protein has any disadvantages. So I'm
not clear whether it is advantageous.

> It seems that apoA-I is responsible for making HDL (high-density
> lipoproteins, or "good cholesterol"). The mutation means that some of
> the proteins it makes will stick together (which "restricts HDL growth",
> hence why people with that mutation have an HDL deficiency), but the
> others will have a free bond that captures free radicals, thus
> preventing the inflammation that causes heart disease.
>
> Now, for one thing point mutations are highly unlikely to be reversed,
> not on the short term at least (consider : there is a very low rate of
> mutation per nucleotide; given there are a lot of nucleotides in the
> genome that still sums up to a lot of mutations overall. But if you're
> looking at the odds of a *specific* mutation happening at a *specific*
> place (i.e. what you need for a specific mutation to be reversed) then
> the probability is... that very low rate of mutation per nucleotide.
> Divided by four, because we're looking for a specific base-pair in this
> case), so it's extremely improbable that apoA-I Milano is a reversion to
> a specific earlier form. Not only that, it's not like we have ancestors
> for whom heart disease due to fatty diets was a *worse* problem than it
> is for us modern humans, that they'd evolve that variant an

> overall. Remember your 20 million mutations over 5 million years between
> our genome and chimpanzees' is based on all the nucleotides, not just
> those that are genes AND are subject to strong selection.

Well firstly I was talking about the human genome as a whole. I assume
estimates of the amount of genetic variability permissible for continued
genetic functionality would have given an indication of how many of the
base pairs should be considered neutral. Secondly I'm not sure what
you mean tiny compared to the size of the human genome overall. As I
understand it, there are only 70-80 thousands genes in the genome. So what
kind of a percentage were you thinking of?

> > Because the speech ones seemed
> > beneficial. This seems to open up questions about what kind of % of the
> > differences should we expect to be of the new mutation type, and what
> > kind of rate per person per generation would they be estimated to be
> > occurring at, and what kind of rate are we observing them at now?
>
> If all those 600 genes were new and beneficial, 600 over 5 million years
> sounds hardly even detectable on a centennial timescale. There are
> probably more in practice; I don't know whether the 600 was "that's how
> many we've positively identified to date" or "that's how many we can
> identify from looking at which parts of the genome underwent selection".

I'm not saying you are, but I'm not thinking that they were suggesting
that only 600 mutations were new and beneficial.

In http://en.wikipedia.org/wiki/Chimpanzee_genome_project it states:
---
About 600 genes have been identified that may have been undergoing strong
positive selection in the human and chimp lineages; many of these genes are
involved in immune system defense against microbial disease (example: granulysin
is protective against Mycobacterium tuberculosis,[3]) or are targeted receptors
of pathogenic microorganisms (example: Glycophorin C and Plasmodium falciparum).
By comparing human and chimp genes to the genes of other mammals, it has been
found that genes coding for transcription factors, such as forkhead-box P2
(FOXP2), have often evolved faster in the human relative to chimp; relatively
small changes in these genes may account for the morphological differences
between humans and chimps. A set of 348 transcription factor genes code for
proteins with an average of about 50 percent more amino acid changes in the
human lineage than in the chimp lineage.
---

Multiple mutations have happened to those same genes. What I'm not sure about
is how many of the mutations happened in these genes, and how many of those
were beneficial.

In http://en.wikipedia.org/wiki/Human_evolutionary_genetics it says:
---
Since mutation rate is relatively constant, roughly one half of these changes occurred in the human lineage. Only a very tiny fraction of those fixed differences gave rise to the different phenotypes of humans and chimpanzees and finding those is a great challenge. The vast majority of the differences are neutral and do not affect the phenotype.[citation needed]
---

I'm not clear on whether every beneficial mutation would count as a change
to phenotype, but if immunology changes for example are, then there is just
the question of much of our genome is neutral.

http://en.wikipedia.org/wiki/Chimpanzee_genome_project mentions that
a chromosome fusing event seems to of happened in the history
of the great apes, and that in humans, the chromsomes 2A and 2B which the other
great apes have are fused to create human chromosome 2. Though between these
two humans seem to have 150,000 additional base pairs. From the description
this seems to me like a DNA sequence could have been copied from chromosome 9
and added to the end of either 2A or 2B before the fusion event. It doesn't
mention how many changes have happened to this 150,000 additional base pair
section. Any idea?

I wonder whether these genes were counted within the 600 genes and if not,
then how many other genes were treated similarly which might
end up being reclassified as beneficial. I'm not saying they will be
reclassified, but from what was said on the website it sounded as though
they weren't sure whether they were beneficial or not.

[Arkalen]

I'm sorry, I find it difficult to understand that paragraph. Are you
wondering about the general odds of an adaptation appearing that
requires 2 mutations or more, neither beneficial on its own ?

I don't know what to say about that beyond what I said in my previous
reply. Is it really relevant to the larger discussion at this point ?

The presence of that sulphur-based residue is directly caused by the
single amino-acid substitution (the replacement amino acid has a sulphur
group on it). Moreover, a different article looked at different versions
of that protein with the substitution happening in different places and
they all functioned differently. So again, we've got the point that
point mutations reversing themselves is vanishingly unlikely. And also,
that the environment in which this mutation provides this immense
benefit didn't exist for our ancestors.

Are you waiting for a comprehensive list of all the versions this
protein has had in our evolutionary history ?

> Though whether it is beneficial or not isn't clear. The article mentions
> that:
> ---
> The trick is to develop a simple pharmaceutical peptide that targets the upregulation of the ABCA1 receptor, exports cholesterol like the conventional protein, and fights arterial wall oxidation like the Milano mutation..
> ---
>
> What I'm not clear on whether not cleaning out the cholesterol of the cells
> as efficiently as the conventional protein has any disadvantages. So I'm
> not clear whether it is advantageous.

The people who have it live to their 90s, eat badly and don't suffer
heart disease. I'm not sure what disadvantages you're picturing that
would stop it being beneficial overall in our environment (because of
course things are only beneficial or not *with respect to the
environment*. No adaptation is beneficial in all circumstances).

At the end of the day you can always say there's some disadvantage we
haven't found yet that makes the mutation non-beneficial, or some
ancestral version we haven't found yet that makes it regressive. But
would you have any reason to think those existed if you didn't have some
other reason to think the mutation isn't new and beneficial ?

Wikipedia said around 20,000 protein-coding genes (though maybe it could
still be what you say when you count regulatory genes and so on). By
"genome as a whole" I meant "all of our nuclear DNA". Most of that
doesn't form genes, and as I said the 20 million or so mutations between
humans and chimpanzee is out of all our DNA, not just the parts that are
genes.
I can't tell what exact percentage of our DNA those 600 genes make up
because I don't know how big those genes are, but it isn't much. But
even if we're only counting genes, 600 out of 70,000 or even 20,000
isn't a lot either.

I'm not sure what it is you want to know; you asked whether I thought
"most of the human genome is made up of beneficial mutations". My
clarification should have made it clear that wasn't what I'd said. I was
saying it seems most of those 600 genes were "new beneficial mutations".
You are aware that there are a lot more genes than 600 in the human
genome, even if you didn't realize most of our DNA isn't genes, so
either way it should have been clear I wasn't saying most of our genome
is made of beneficial mutations.

>
>>> Because the speech ones seemed
>>> beneficial. This seems to open up questions about what kind of % of the
>>> differences should we expect to be of the new mutation type, and what
>>> kind of rate per person per generation would they be estimated to be
>>> occurring at, and what kind of rate are we observing them at now?
>>
>> If all those 600 genes were new and beneficial, 600 over 5 million years
>> sounds hardly even detectable on a centennial timescale. There are
>> probably more in practice; I don't know whether the 600 was "that's how
>> many we've positively identified to date" or "that's how many we can
>> identify from looking at which parts of the genome underwent selection".
>
> I'm not saying you are, but I'm not thinking that they were suggesting
> that only 600 mutations were new and beneficial.

I've found the paper in question, it's the first citation in the
Wikipedia page ("Initial sequence of the chimpanzee genome and
comparison with the human genome") and the PDF is directly linked to so
it looks like it might be open access.

In it they say they looked at 13,454 genes that obviously corresponded
to each other in the human and chimpanzee genomes. That's a "large
fraction of the entire complement of human genes"; sadly they don't say
how large a fraction. The 585 genes they identified as having
experienced strong positive selection are out of those 13,454; the way
they do this is look at which genes have accumulated amino acid
substitutions faster than the neutral substitution rate would predict.
However they also say that these higher subtitution rates could be
expected to occur by chance in "at least 263 cases", so the actual
number could be much lower.

In any case if you upscale it to all human genes I doubt you'd get
higher than, say, 6000 new beneficial genes.

>
> In http://en.wikipedia.org/wiki/Chimpanzee_genome_project it states:
> ---
> About 600 genes have been identified that may have been undergoing strong
> positive selection in the human and chimp lineages; many of these genes are
> involved in immune system defense against microbial disease (example: granulysin
> is protective against Mycobacterium tuberculosis,[3]) or are targeted receptors
> of pathogenic microorganisms (example: Glycophorin C and Plasmodium falciparum).
> By comparing human and chimp genes to the genes of other mammals, it has been
> found that genes coding for transcription factors, such as forkhead-box P2
> (FOXP2), have often evolved faster in the human relative to chimp; relatively
> small changes in these genes may account for the morphological differences
> between humans and chimps. A set of 348 transcription factor genes code for
> proteins with an average of about 50 percent more amino acid changes in the
> human lineage than in the chimp lineage.
> ---
>
> Multiple mutations have happened to those same genes. What I'm not sure about
> is how many of the mutations happened in these genes, and how many of those
> were beneficial.

According to the same paper the proportion of amino acid changes
attributable to positive selection according to their analysis is quite
low. A previous analysis (looking at fewer genes and Old World monkeys
instead of chimpanzees) had estimated about 35% by comparing the
differences within humans with the differences between humans and
monkeys, but in this analysis when they calculated the differences
within humans and the differences between humans and chimpanzees they
didn't find a statistically significant difference between the two; i.e.
they can't give a number for the proportion of amino acid changes
attributable to positive selection, but it's much lower than 35% (the
error bars encompass 0, basically)

>
> In http://en.wikipedia.org/wiki/Human_evolutionary_genetics it says:
> ---
> Since mutation rate is relatively constant, roughly one half of these changes occurred in the human lineage. Only a very tiny fraction of those fixed differences gave rise to the different phenotypes of humans and chimpanzees and finding those is a great challenge. The vast majority of the differences are neutral and do not affect the phenotype.[citation needed]
> ---
>
> I'm not clear on whether every beneficial mutation would count as a change
> to phenotype, but if immunology changes for example are, then there is just
> the question of much of our genome is neutral.

By definition a beneficial mutation has to change the phenotype,
otherwise it would be neutral.

>
> http://en.wikipedia.org/wiki/Chimpanzee_genome_project mentions that
> a chromosome fusing event seems to of happened in the history
> of the great apes, and that in humans, the chromsomes 2A and 2B which the other
> great apes have are fused to create human chromosome 2. Though between these
> two humans seem to have 150,000 additional base pairs. From the description
> this seems to me like a DNA sequence could have been copied from chromosome 9
> and added to the end of either 2A or 2B before the fusion event. It doesn't
> mention how many changes have happened to this 150,000 additional base pair
> section. Any idea?
>
> I wonder whether these genes were counted within the 600 genes and if not,
> then how many other genes were treated similarly which might
> end up being reclassified as beneficial. I'm not saying they will be
> reclassified, but from what was said on the website it sounded as though
> they weren't sure whether they were beneficial or not.
>

Apparently one of the six regions that experienced a strong selective
sweep was on Chromosome 2, including two genes and part of a third. You
can look up GTDC1, ZFHX1B and ARHGAP15; I wasn't able to figure out
whether they were part of the original chromosomes or the bits that came
from others. Possibly part of the original, given GTDC1 and ARHGAP15
seem to be on Chromosome 2 but not another one.

At the end of the day I don't know what it is you're trying to find out
here. Even if humans had 80,000 genes and all had evolved since we
diverged from chimpanzees that would still be less than two a century.
If we had 6000 new beneficial mutations since chimpanzees that would be
one every 800 years. Considering the human population was much smaller
in the past (so fixation would happen much faster) that doesn't look
unreasonable to me.

[someone]

On Thursday, 20 June 2013 17:47:09 UTC+1, Arkalen wrote:
> On 20/06/13 11:50, someone wrote:
> > On Wednesday, 19 June 2013 20:59:02 UTC+1, Arkalen wrote:
> >> On 19/06/13 19:07, someone wrote:

[snip]

> > Also I now realise that not only did I assume only one candidate, but
> > I assumed it would be more than 1 mutation away from just encoding
> > whatever would be needed to be encoded for something to react with what
> > seemed like a neutral mutation which nothing else which had been developed
> > seemed to have any particular reaction to, and to do so to create a new
> > beneficial mutation. Remove that assumption and then I wouldn't expect
> > the neutral mutation to be more likely corrupted. Do you know of a
> > reason to remove that assumption? The reason I had it was because I saw no
> > reason for any bias towards encoding whatever it was that would have been
> > beneficial reacting directly or indirectly with the neutral mutation that got fixed.
> I'm sorry, I find it difficult to understand that paragraph. Are you
> wondering about the general odds of an adaptation appearing that
> requires 2 mutations or more, neither beneficial on its own ?
> I don't know what to say about that beyond what I said in my previous
> reply. Is it really relevant to the larger discussion at this point ?

The larger discussion of is there anything unexpected in the mutation
frequencies? Well it might be relevant to discussions of expected
frequencies of new beneficial mutations.

[snip - I am snipping more because of the need to keep removing all these
extra lines]

> > I read that it said:
> > ---
> > The mutated protein, however, boasts an antioxidant in the form of a sulfur-based residue that mops up unpaired electrons and prevents them from triggering arterial inflammation
> > ---
> > If a protein which boasted such an antioxidant wasn't present before in the
> > evolutionary history of their genome then it would seem like a new mutation.
> The presence of that sulphur-based residue is directly caused by the
> single amino-acid substitution (the replacement amino acid has a sulphur
> group on it). Moreover, a different article looked at different versions
> of that protein with the substitution happening in different places and
> they all functioned differently. So again, we've got the point that
> point mutations reversing themselves is vanishingly unlikely. And also,
> that the environment in which this mutation provides this immense
> benefit didn't exist for our ancestors.
>
> Are you waiting for a comprehensive list of all the versions this
> protein has had in our evolutionary history ?
>

No why should I be? You said it was distinct functionality to do with the
mutated base pair, and I had said: "If a protein which boasted such an

antioxidant wasn't present before in the evolutionary history of their genome
then it would seem like a new mutation."

>

> > Though whether it is beneficial or not isn't clear. The article mentions
> > that:
> > ---
> > The trick is to develop a simple pharmaceutical peptide that targets the upregulation of the ABCA1 receptor, exports cholesterol like the conventional protein, and fights arterial wall oxidation like the Milano mutation..
> > ---
> > What I'm not clear on whether not cleaning out the cholesterol of the cells
> > as efficiently as the conventional protein has any disadvantages. So I'm
> > not clear whether it is advantageous.
>
> The people who have it live to their 90s, eat badly and don't suffer
> heart disease. I'm not sure what disadvantages you're picturing that
> would stop it being beneficial overall in our environment (because of
> course things are only beneficial or not *with respect to the
> environment*. No adaptation is beneficial in all circumstances).
>
> At the end of the day you can always say there's some disadvantage we
> haven't found yet that makes the mutation non-beneficial, or some
> ancestral version we haven't found yet that makes it regressive. But
> would you have any reason to think those existed if you didn't have some
> other reason to think the mutation isn't new and beneficial ?
>
>

The reason I mentioned it was that the doctors were looking to combine the
benefits of the normal gene and the mutated gene. Suggesting that the normal
gene had some benefit they felt was still useful. When checking whether
it is beneficial or not why not compare between two healthy people that eat
well?

[snip]

>
> > Well firstly I was talking about the human genome as a whole. I assume
> > estimates of the amount of genetic variability permissible for continued
> > genetic functionality would have given an indication of how many of the
> > base pairs should be considered neutral. Secondly I'm not sure what
> > you mean tiny compared to the size of the human genome overall. As I
> > understand it, there are only 70-80 thousands genes in the genome. So what
> > kind of a percentage were you thinking of?
>
> Wikipedia said around 20,000 protein-coding genes (though maybe it could
> still be what you say when you count regulatory genes and so on). By
> "genome as a whole" I meant "all of our nuclear DNA". Most of that
> doesn't form genes, and as I said the 20 million or so mutations between
> humans and chimpanzee is out of all our DNA, not just the parts that are
> genes.
>

I got the figures from here http://www.ncbi.nlm.nih.gov/books/NBK7587/
though I hadn't read it, they just showed on google search. I can see
similar figures in section 7.2.1 though.

> I can't tell what exact percentage of our DNA those 600 genes make up
> because I don't know how big those genes are, but it isn't much. But
> even if we're only counting genes, 600 out of 70,000 or even 20,000
> isn't a lot either.
>
> I'm not sure what it is you want to know; you asked whether I thought
> "most of the human genome is made up of beneficial mutations". My
> clarification should have made it clear that wasn't what I'd said. I was
> saying it seems most of those 600 genes were "new beneficial mutations".
> You are aware that there are a lot more genes than 600 in the human
> genome, even if you didn't realize most of our DNA isn't genes, so
> either way it should have been clear I wasn't saying most of our genome
> is made of beneficial mutations.
>

They mentioned in the article that:
---
The above values suggest that the genes in the nuclear genome represent about
99.95% of the total number of cellular genes. If the average size of a human
nuclear gene, including introns, is taken to be about 10�15 kb, this would mean
that if the genes did not show overlaps, the total nuclear DNA occupied by genes
would be about 70 000 � (10�15) kb or about 700�1050 Mb which corresponds
roughly to about 25�35% of the genome. As the vast majority of nuclear genes
encode polypeptides and the coding sequence required for an average size human
polypeptide is taken to be about 500�600 codons, that is 1.5�1.8 kb, only about
3% of the nuclear genome (80�100 Mb of the 3300 Mb) would be expected to have a
coding function.
---

So if the 70,000 genes took up say 30% of the genome, then I'm assuming
600 genes would be about 0.26% of the total genome.

I'm not sure what you mean by 'it seems most of those 600 genes were "new beneficial mutations"'.

What I meant was was that at least some of the genes had been mutated more than once.

But if the 600 genes had 10 or more beneficial mutations on each for example,
I'm not saying they did, I'm just making that up for the example, then we
could look at how much of the total DNA that occurred in. Let's say it was a
quarter of a percent. Was the frequency of beneficial mutation in that quarter
of a percent statistically surprising? Also taking into account the smaller
population http://www.ncdc.noaa.gov/paleo/ctl/100k.html states that around
10,000 years ago the world population was about 5 million. Therefore if
we assume the chances of a person having a new beneficial mutation is the same
now as it was then, given the population is now over 5 billion, new beneficial
mutations would be expected to be a thousand times more frequent now than
then. Perhaps even considerably more, since the population might have averaged
as around 50,000 for a considerable period of the 5 million years for example.

And what if some of those genes in the 150,000 base pairs section between the
fusing of the chromosomes are beneficial, and have undergone several mutations
to be so. Even if not, if they aren't included in the 600 genes, then there
might also be other genes whose function or lack of it is unknown, and might
be shown to be beneficial. At the moment though, I am just interested with how
many changes happened to the 600 genes in question, just to get an estimate on
how many new beneficial mutations a person would have been expected to get a
year in order to fit the evidence. Because once we have that, we can check with
the current rate of new beneficial mutation, and see whether it matches.

I pretty much assume that somebody somewhere has already done this, and that
there is nothing of interest, but I thought I'd check, and am surprised that
as far as I have seen it doesn't seem to be directly addressed in the wiki
literature.

[Arkalen]

Oh, I'm sorry, I misinterpreted ! I thought you were saying you weren't
convinced such an antioxidant wasn't present before.

>
>>
>>> Though whether it is beneficial or not isn't clear. The article mentions
>>> that:
>>> ---
>>> The trick is to develop a simple pharmaceutical peptide that targets the upregulation of the ABCA1 receptor, exports cholesterol like the conventional protein, and fights arterial wall oxidation like the Milano mutation..
>>> ---
>>> What I'm not clear on whether not cleaning out the cholesterol of the cells
>>> as efficiently as the conventional protein has any disadvantages. So I'm
>>> not clear whether it is advantageous.
>>
>> The people who have it live to their 90s, eat badly and don't suffer
>> heart disease. I'm not sure what disadvantages you're picturing that
>> would stop it being beneficial overall in our environment (because of
>> course things are only beneficial or not *with respect to the
>> environment*. No adaptation is beneficial in all circumstances).
>>
>> At the end of the day you can always say there's some disadvantage we
>> haven't found yet that makes the mutation non-beneficial, or some
>> ancestral version we haven't found yet that makes it regressive. But
>> would you have any reason to think those existed if you didn't have some
>> other reason to think the mutation isn't new and beneficial ?
>>
>>
>
> The reason I mentioned it was that the doctors were looking to combine the
> benefits of the normal gene and the mutated gene. Suggesting that the normal
> gene had some benefit they felt was still useful. When checking whether
> it is beneficial or not why not compare between two healthy people that eat
> well?

Actually I'm not sure why I'd said the people involved eat badly; they
have low HDL but I haven't found the source where I'd seen that (if I
didn't make it up entirely). Quite the opposite, from the original paper
it seems the person who'd first been found with the gene had been on
quite a few diets, and the paper looking at the village the mutation is
thought to have originated in (both papers linked at in the Wikipedia
citations) talks about the villagers all having the same diets, yet the
carriers historically had "relatively long lives" for the time and
period. So diet appears to have been controlled for.

I don't think what the doctors are trying to do here is that relevant;
for one thing, they've got a requirement to do no harm so they really
need to be certain of what potential side-effects are; that's different
from determining whether a mutation is beneficial or not, where all it
needs is to be beneficial overall.
If doctors were examining a mutation that made the brain larger they
wouldn't prescribe it either because it would increase the risk of death
in childbirth for both the mother and the child. Which it does. Yet our
intelligence (strongly related to our brain size) is generally
considered a beneficial adaptation despite that grave drawback. Would
you disagree ?

Also, doctors can't change a person's genome entirely and most people
have normal HDL, so any therapy that applied to everyone would take that
into account.

Finally, the same paper that looks at the history of the mutation says
all the carriers found are heterozygous, which suggests they have normal
HDL proteins too, just not as many as people homozygous for the
wild-type gene.

Thank you for looking up the numbers and linking to that book, it's very
informative.

>
> I'm not sure what you mean by 'it seems most of those 600 genes were "new beneficial mutations"'.

I'm going back to a few exchanges earlier, specifically this paragraph:

"Hey, is it me or is the Wikipedia page on the chimpanzee genome project
more detailed than it was last time I looked ? According to it they
identified 600 genes that underwent strong selection in humans and
chimpanzees; we can call those the "beneficial mutations". They don't
give numbers on which genes do what, but nearly all those they mention
seem like they'd be non-regressions by your definition."

To which your reply asked if I was "saying that most of the human genome
is made up of new beneficial mutations". That was obviously not the case
but on re-reading I understand your reply better now, given you'd
originally asked me how much of the human genome I was thinking
displayed "new beneficial mutations".

Of course they were, but I don't know what you think that means. Or
wait, were you wondering how many of the mutations in a given gene were
beneficial and how many were neutral ? Is this another way of looking at
the odds of beneficial mutations happening on top of a
necessary-but-neutral mutation ?

Mutations are more frequent but the number of mutations *per person* is
pretty much the same, and the higher number of people means it will take
more time for a mutation to spread to a point we're likely to notice it.

Or put another way, we're not looking at the genomes of 7 billion
people. The odds of us finding new mutations depends on how many
mutations occur *among the people we're looking at*, not the whole
population.

>
> And what if some of those genes in the 150,000 base pairs section between the
> fusing of the chromosomes are beneficial, and have undergone several mutations
> to be so. Even if not, if they aren't included in the 600 genes,

They're as likely to be included in the 600 genes as any other section
of the genome is, insofar as chimpanzee orthologs could be identified on
them and given they've been identified as copies of other chromosomes I
don't see why not.

> then there
> might also be other genes whose function or lack of it is unknown, and might
> be shown to be beneficial.

As I explained, there obviously are given those 600 genes were out of
13,000 they found that could easily be compared with the chimpanzee
versions, not the whole genome. No need to invoke chromosome 2 for that.

> At the moment though, I am just interested with how
> many changes happened to the 600 genes in question, just to get an estimate on
> how many new beneficial mutations a person would have been expected to get a
> year in order to fit the evidence. Because once we have that, we can check with
> the current rate of new beneficial mutation, and see whether it matches.
>
> I pretty much assume that somebody somewhere has already done this, and that
> there is nothing of interest, but I thought I'd check, and am surprised that
> as far as I have seen it doesn't seem to be directly addressed in the wiki
> literature.
>

I highly doubt anybody has done this because we don't know enough to do
so. We don't have a "current rate of new beneficial mutations". We don't
know enough about the genome to identify a mutation as beneficial unless
it has a large and obvious effect on the phenotype (which wouldn't be
the case of most beneficial mutations), and we aren't looking at a large
proportion of all genomes in the first place. The way fixed mutations
are identified as beneficial is in looking at whether they experienced
positive selection; this can't be done on a mutation that's only just
appeared or has only been around for a few generations. And it's more
difficult to do the smaller the section of genome you're looking at is,
so there's no way of telling a beneficial mutation from a neutral one on
a single gene with that method.

And as for the past rate of beneficial mutations, surely you've noticed
with all the research we've done that all of this knowledge we have
about the evolution of the human genome is extremely recent. The paper
on the chimpanzee genome project is from 2005. Our knowledge is
improving every day but we aren't nearly at the point we can give the
precise numbers you're asking for.

[someone]

On Friday, 21 June 2013 14:01:42 UTC+1, Arkalen wrote:
> On 21/06/13 01:24, someone wrote:
> > On Thursday, 20 June 2013 17:47:09 UTC+1, Arkalen wrote:
> >> On 20/06/13 11:50, someone wrote:
> >>> On Wednesday, 19 June 2013 20:59:02 UTC+1, Arkalen wrote:
> >>>> On 19/06/13 19:07, someone wrote:
[snip]
>

I would only expect doctors to perscribe such drugs if the baby's brain
was expected to be much smaller than normal. I agree that our larger
brain size is generally considered a beneficial adaptation.

>
> Also, doctors can't change a person's genome entirely and most people
> have normal HDL, so any therapy that applied to everyone would take that
> into account.
>
> Finally, the same paper that looks at the history of the mutation says
> all the carriers found are heterozygous, which suggests they have normal
> HDL proteins too, just not as many as people homozygous for the
> wild-type gene.
>

Is the mutated gene not dominant?

With regards to the doctors, I assumed the companies were going to make
drugs to deliver the same kind of effects that the genes do. Were you thinking
that they were going to change the DNA with some sort of viral carrier?

[snip]

>
> > What I meant was was that at least some of the genes had been mutated more than once.
>
>
> Of course they were, but I don't know what you think that means. Or
> wait, were you wondering how many of the mutations in a given gene were
> beneficial and how many were neutral ? Is this another way of looking at
> the odds of beneficial mutations happening on top of a
> necessary-but-neutral mutation ?
>

I thought I'd been quite clear on what I was considering a beneficial mutation
to be. Are you using the term in a different way, because I thought you could
work with how I was using it.

[snip]

While how many people looked at is a factor, an estimate of amount of new
beneficial mutations could be estimated based on sample size. Since if new
beneficial mutations were frequent, then there would be expected to be
blooms of such mutations across the population, and a person might show
several, even though the ones they showed happened a thousand to two thousand
years ago for example. The example was just made up by the way, I don't know
how many new beneficial mutations that occurred in the past couple of thousand
years would be expected given the previous 5 million year rate per person.

The new beneficial mutation rate per person (which as you say would be expected
to be pretty much the same) could be estimate from the number of beneficial
mutations found on the 600 genes in question, and an estimate of the
average population up to say 50,000 years ago. This method would be expected
to slightly overestimate the number of mutations, and if they seemed unexpectedly high compared to what is observed, then If this was found
to be considerably higher than what is currently being observed, then further
estimates with the amount of the mutations that would have taken place from
50,000 years ago to the present being taken into account along with the
increasing population of the period.

> > And what if some of those genes in the 150,000 base pairs section between the
> > fusing of the chromosomes are beneficial, and have undergone several mutations
> > to be so. Even if not, if they aren't included in the 600 genes,
>
> They're as likely to be included in the 600 genes as any other section
> of the genome is, insofar as chimpanzee orthologs could be identified on
> them and given they've been identified as copies of other chromosomes I
> don't see why not.
>
>

Yes, but, as I'm understanding it (and I could easily be misunderstanding it),
these genes aren't known to be beneficial, and the 600 genes are genes where
the mutation rate couldn't be explained by genetic drift and so some are assumed
to be positively selected. The frequency of mutations on the 150,000 base pair
section might not be so high that it couldn't be explained by genetic drift. So
while it could for all I know contain beneficial mutations, they aren't genes
tagged for needing beneficial mutations to explain the mutation rate.

>
> > then there
> > might also be other genes whose function or lack of it is unknown, and might
> > be shown to be beneficial.
>
> As I explained, there obviously are given those 600 genes were out of
> 13,000 they found that could easily be compared with the chimpanzee
> versions, not the whole genome. No need to invoke chromosome 2 for that.
>

So how many of the 13,000 would you be guessing at contained beneficial
mutations?

Chromosome 2 was only mentioned in case there were benefical mutations
on it in the additional 150,000 base pairs. If any of those genes
did turn out to contain beneficial mutations then I was just curious as
to whether given the rate, the amount over the amount of base pairs
would be surprising. I'm not saying it would, but out of curiousity I'd
like to know the answers.

> > At the moment though, I am just interested with how
> > many changes happened to the 600 genes in question, just to get an estimate on
> > how many new beneficial mutations a person would have been expected to get a
> > year in order to fit the evidence. Because once we have that, we can check with
> > the current rate of new beneficial mutation, and see whether it matches.
> > I pretty much assume that somebody somewhere has already done this, and that
> > there is nothing of interest, but I thought I'd check, and am surprised that
> > as far as I have seen it doesn't seem to be directly addressed in the wiki
> > literature.
>
> I highly doubt anybody has done this because we don't know enough to do
> so. We don't have a "current rate of new beneficial mutations". We don't
> know enough about the genome to identify a mutation as beneficial unless
> it has a large and obvious effect on the phenotype (which wouldn't be
> the case of most beneficial mutations), and we aren't looking at a large
> proportion of all genomes in the first place. The way fixed mutations
> are identified as beneficial is in looking at whether they experienced
> positive selection; this can't be done on a mutation that's only just
> appeared or has only been around for a few generations. And it's more
> difficult to do the smaller the section of genome you're looking at is,
> so there's no way of telling a beneficial mutation from a neutral one on
> a single gene with that method.

Maybe ones that have been around for 500 years, or 1000 or 5000 perhaps
might be more noticeable.

Let's imagine that if a certain environmental factor reached a certain level
that there would be health issues for the organism. Let's then say that either
a single mutation or a series of mutations had produced a new regulation
method which worked to lower the level of the factor when higher levels of it
were reached. Could such mutation(s) not be considered as beneficial if no disadvantageous side effects could be discerned?

>
> And as for the past rate of beneficial mutations, surely you've noticed
> with all the research we've done that all of this knowledge we have
> about the evolution of the human genome is extremely recent. The paper
> on the chimpanzee genome project is from 2005. Our knowledge is
> improving every day but we aren't nearly at the point we can give the
> precise numbers you're asking for.

I just thought that there might be the latest estimate, and by that I mean
what they estimate the rate to have needed to have been from what they know
so far.

[Arkalen]

Another way in which doctors differ from natural selection. Natural
selection doesn't know or care about "normal". Doctors do.

> I agree that our larger
> brain size is generally considered a beneficial adaptation.
>
>>
>> Also, doctors can't change a person's genome entirely and most people
>> have normal HDL, so any therapy that applied to everyone would take that
>> into account.
>>
>> Finally, the same paper that looks at the history of the mutation says
>> all the carriers found are heterozygous, which suggests they have normal
>> HDL proteins too, just not as many as people homozygous for the
>> wild-type gene.
>>
>
> Is the mutated gene not dominant?

IIRC whether a gene is dominant or not is about its final effect on the
phenotype, it isn't some gene-level mechanism where dominant genes turn
off recessive ones for example. So whether a gene is dominant or not
doesn't say anything about which proteins are being created at the
molecular level. (to be fair I'm not sure I was right in assuming
heterozygous individual necessarily produce both proteins either)

>
> With regards to the doctors, I assumed the companies were going to make
> drugs to deliver the same kind of effects that the genes do. Were you thinking
> that they were going to change the DNA with some sort of viral carrier?

No, I was confused about your statement that doctors are trying to

combine the benefits of the normal gene and the mutated gene.

>

> [snip]
>>
>>> What I meant was was that at least some of the genes had been mutated more than once.
>>
>>
>> Of course they were, but I don't know what you think that means. Or
>> wait, were you wondering how many of the mutations in a given gene were
>> beneficial and how many were neutral ? Is this another way of looking at
>> the odds of beneficial mutations happening on top of a
>> necessary-but-neutral mutation ?
>>
>
> I thought I'd been quite clear on what I was considering a beneficial mutation
> to be. Are you using the term in a different way, because I thought you could
> work with how I was using it.

I wasn't asking about your definition of beneficial mutation, I just
didn't know why you thought it's relevant that some genes were mutated
more than once. I now think you were wondering what the proportion of
beneficial vs neutral mutations in those genes as a prerequisite before
calculating the proportion for the whole genome, would that be accurate ?

We do observe many beneficial mutations that spread within the last
10,000 years, mostly linked to diet and resistance to disease (all of
those lactose tolerance genes and so forth).

I don't think we have enough systematic data to calculate an overall
rate from that. By "systematic" I mean "look at the whole genome or some
statistically-valid subset thereof and find the beneficial mutations",
not "how hey look this here is a beneficial mutation", which AFAIK is
how things like lactose-tolerance genes were found.

> The example was just made up by the way, I don't know
> how many new beneficial mutations that occurred in the past couple of thousand
> years would be expected given the previous 5 million year rate per person.
>
> The new beneficial mutation rate per person (which as you say would be expected
> to be pretty much the same)

> could be estimate from the number of beneficial
> mutations found on the 600 genes in question, and an estimate of the
> average population up to say 50,000 years ago. This method would be expected
> to slightly overestimate the number of mutations, and if they seemed unexpectedly high compared to what is observed, then If this was found
> to be considerably higher than what is currently being observed, then further
> estimates with the amount of the mutations that would have taken place from
> 50,000 years ago to the present being taken into account along with the
> increasing population of the period.

Of course note that when we're looking at beneficial mutations that
appeared in the last few thousand years we're already no longer looking
at the rate at which beneficial mutations appear but also at their rate
of spread which also depends on population dynamics.

Also, a point of pedantry : the number of *mutations* per person could
be expected to be pretty much the same (although even there, with
increased atmospheric radioactivity, industrial chemicals and stuff who
knows), but whether a mutation is beneficial or not depends on the
environment, so the rate of appearance of beneficial mutations could
vary with the environment or genome characteristics. For example, maybe
an organism outside its optimal environment would have a higher rate of
beneficial mutations than one who's in a stable environment they've been
adapted to for millions of years.
But I don't see many reasons to think proto-humans and modern humans
would have had very different rates of beneficial mutations.

>
>>> And what if some of those genes in the 150,000 base pairs section between the
>>> fusing of the chromosomes are beneficial, and have undergone several mutations
>>> to be so. Even if not, if they aren't included in the 600 genes,
>>
>> They're as likely to be included in the 600 genes as any other section
>> of the genome is, insofar as chimpanzee orthologs could be identified on
>> them and given they've been identified as copies of other chromosomes I
>> don't see why not.
>>
>>
>
> Yes, but, as I'm understanding it (and I could easily be misunderstanding it),
> these genes aren't known to be beneficial, and the 600 genes are genes where
> the mutation rate couldn't be explained by genetic drift and so some are assumed
> to be positively selected. The frequency of mutations on the 150,000 base pair
> section might not be so high that it couldn't be explained by genetic drift. So
> while it could for all I know contain beneficial mutations, they aren't genes
> tagged for needing beneficial mutations to explain the mutation rate.
>

If a section of genome hasn't been subject to much positive selection at
a given time, in what sense would mutations that occurred in that gene
at that time be "beneficial" ? They clearly didn't affect the organism's
reproductive success much. If you're thinking of mutations that are only
slightly beneficial, then without a selective signature of some sort I
don't know how we could detect them.

Looking at the specific effects of each gene isn't something we can do
on a systematic basis, and even those genes whose effects we *do* know,
we don't know every single side-effect they could have (cf how you're
still not convinced apoA-1 Milano doesn't have SOME deleterious
side-effect. And you're right; our understanding of human physiology
isn't even close to detailed enough to completely it out. And that's one
gene that we've been studying for thirty years).

Looking at whether genes have undergone positive selection is a
convenient way of getting around that problem; it's a bit like having to
calculate exactly where a piece of paper will fall when you drop it from
a second-floor window. Calculating that in advance is so hard as to be
near-impossible; "calculating" it after the fact from looking at where
the piece of paper actually fell is both much easier and much more
accurate than any a priori calculation you could have done.

Same with genes. Finding out 1) all of a mutation's benefits and
drawbacks and 2) whether or not those benefits will outweigh those
drawbacks in a given environment is hard. But if we know a mutation
experienced positive selection (or negative for that matter) then we
know that its benefits DID outweigh its drawbacks (or vice-versa) on
average in that period.

>
>>
>>> then there
>>> might also be other genes whose function or lack of it is unknown, and might
>>> be shown to be beneficial.
>>
>> As I explained, there obviously are given those 600 genes were out of
>> 13,000 they found that could easily be compared with the chimpanzee
>> versions, not the whole genome. No need to invoke chromosome 2 for that.
>>
>
> So how many of the 13,000 would you be guessing at contained beneficial
> mutations?

Either you already know my answer to that (around 600!) because you know
what paper and what Wikipedia page I'm getting all my information from
on this so I don't know why you're asking, or you're asking for
information beyond what's in the Wikipedia page and the paper involved
in which case I need to better understand what you want because I have
no idea what it is I need to look up here.

Are you asking me how many beneficial mutations there are in the genes
that weren't strongly selected for ?

Do you have a suggestion for how that number could even be calculated
given our current understanding of genetics ?

Sure. There are many examples of such mutations, but as I said earlier I
don't know if we have enough information to derive an overall rate from
that. I'll try and find out more though.

>
>>
>> And as for the past rate of beneficial mutations, surely you've noticed
>> with all the research we've done that all of this knowledge we have
>> about the evolution of the human genome is extremely recent. The paper
>> on the chimpanzee genome project is from 2005. Our knowledge is
>> improving every day but we aren't nearly at the point we can give the
>> precise numbers you're asking for.
>
> I just thought that there might be the latest estimate, and by that I mean
> what they estimate the rate to have needed to have been from what they know
> so far.
>

We don't know the rates to that level of precision that I know of.

I'm going to help you though : if, out of all the information we DO
have, there was a surprising inconsistency between the amount of
mutations required to differentiate from our common ancestor with
chimpanzees and the amount of mutations appearing now or expected to
have appeared then, *this would be an active area of research*. You'd
have articles with titles like "Resolving the mutation rate paradox" or
"Has the mutation rate changed ? An inflationary theory of human origins".

I've never heard of this being a particular research question (and a
cursory Google scholar search yields nothing) so it looks as if as all
the information geneticists and anthropologists have on the mutation
rate in human evolution is internally consistent.

[Darwin123]

On Friday, June 21, 2013 9:01:42 AM UTC-4, Arkalen wrote:
> On 21/06/13 01:24, someone wrote:
>
> > On Thursday, 20 June 2013 17:47:09 UTC+1, Arkalen wrote:
>
> >> On 20/06/13 11:50, someone wrote:
>
> >>> On Wednesday, 19 June 2013 20:59:02 UTC+1, Arkalen wrote:
>
> >>>> On 19/06/13 19:07, someone wrote:
>
> > [snip]
>
> >
>

-clip-

>
> Actually I'm not sure why I'd said the people involved eat badly; they
>
> have low HDL but I haven't found the source where I'd seen that (if I
>
> didn't make it up entirely). Quite the opposite, from the original paper
>
> it seems the person who'd first been found with the gene had been on
>
> quite a few diets, and the paper looking at the village the mutation is
>
> thought to have originated in (both papers linked at in the Wikipedia
>
> citations) talks about the villagers all having the same diets, yet the
>
> carriers historically had "relatively long lives" for the time and
>
> period. So diet appears to have been controlled for.

The statistical methods that they used don't separate anatomy
from behavior. Maybe the mutation caused a change in the behavior
so that the person ate better.

Remember that behavior is part of phenotype. Sure,
environment and training go into behavior. Environment
and training effect every part of the phenotype. Maybe the
gene affected the persons sweet tooth, their sensitivity to
insulin, or what have you. Or maybe the mutation simply made
the person smarter. A more intelligent person is more likely
to eat right.

The mutated gene may make the person far more intelligent
but decrease the bodies response to insulin. The first is a
"benefit" because the person then eats right. However, the second
is a "disadvantage" because it causes diabetes. If the gene
spreads, then one knows that the gene is on average "beneficial".
However, the rate of spreading doesn't tell you why it
was "on average" beneficial.

The concept of "benefit" is ambiguous. This is because there
is no fundamental chemistry or physics behind it.



This is part of the problem with "statistics only" studies. The
statistical methods determine correlation, not cause. A gene can have
multiple effects on the organism. Statistical methods only determine
the average effect over long time periods. The "average effect" is
real. However, the chemical and physical processes are hidden by the
"averaging process".

You seem to think that there is a chemical or physical test to
determine that a gene is "beneficial". Maybe the problem is that the
authors of that article did not include the words "on average". They
determined which mutations were beneficial "on average", but not which
mutations were "absolutely beneficial."

Personally, I doubt that there are any mutations that are "absolutely
beneficial". Evolution depends on mutations that are "beneficial on average".

-clip-

[Arkalen]

On 24/06/13 23:44, Darwin123 wrote:
> On Friday, June 21, 2013 9:01:42 AM UTC-4, Arkalen wrote:
>> On 21/06/13 01:24, someone wrote:
>>
>>> On Thursday, 20 June 2013 17:47:09 UTC+1, Arkalen wrote:
>>
>>>> On 20/06/13 11:50, someone wrote:
>>
>>>>> On Wednesday, 19 June 2013 20:59:02 UTC+1, Arkalen wrote:
>>
>>>>>> On 19/06/13 19:07, someone wrote:
>>
>>> [snip]
>>
>>>
>>
> -clip-
>>
>> Actually I'm not sure why I'd said the people involved eat badly; they
>>
>> have low HDL but I haven't found the source where I'd seen that (if I
>>
>> didn't make it up entirely). Quite the opposite, from the original paper
>>
>> it seems the person who'd first been found with the gene had been on
>>
>> quite a few diets, and the paper looking at the village the mutation is
>>
>> thought to have originated in (both papers linked at in the Wikipedia
>>
>> citations) talks about the villagers all having the same diets, yet the
>>
>> carriers historically had "relatively long lives" for the time and
>>
>> period. So diet appears to have been controlled for.
> The statistical methods that they used don't separate anatomy
> from behavior. Maybe the mutation caused a change in the behavior
> so that the person ate better.

1) That's even more unjustified than assuming the mutation has
deleterious side-effects. Sure, the mutation could do *anything*, but
from the evidence we have some effects are vastly more likely than
others. Like, say, a direct anti-inflammatory effect on the arteries.

2) I don't think they used statistical methods to control for behaviour
(my abuse of language there I guess). They just observed all the
villagers had the same basic diet. That accounts for large-scale
differences in diet (all McDonalds all the time vs Mediterranean).
Again, small differences could have an effect but the point of this
mutation is that people have little arteriosclerosis despite having
extremely low HDL. No variation in diet will make that happen, let alone
a subtle one.

>
> Remember that behavior is part of phenotype. Sure,
> environment and training go into behavior. Environment
> and training effect every part of the phenotype. Maybe the
> gene affected the persons sweet tooth, their sensitivity to
> insulin, or what have you. Or maybe the mutation simply made
> the person smarter. A more intelligent person is more likely
> to eat right.
>
> The mutated gene may make the person far more intelligent
> but decrease the bodies response to insulin. The first is a
> "benefit" because the person then eats right. However, the second
> is a "disadvantage" because it causes diabetes. If the gene
> spreads, then one knows that the gene is on average "beneficial".
> However, the rate of spreading doesn't tell you why it
> was "on average" beneficial.
>
> The concept of "benefit" is ambiguous. This is because there
> is no fundamental chemistry or physics behind it.
>

That's true.

>
> This is part of the problem with "statistics only" studies. The
> statistical methods determine correlation, not cause. A gene can have
> multiple effects on the organism. Statistical methods only determine
> the average effect over long time periods. The "average effect" is
> real. However, the chemical and physical processes are hidden by the
> "averaging process".
>
> You seem to think that there is a chemical or physical test to
> determine that a gene is "beneficial". Maybe the problem is that the
> authors of that article did not include the words "on average". They
> determined which mutations were beneficial "on average", but not which
> mutations were "absolutely beneficial."

I don't know where you got the idea I think that. I've repeatedly
pointed out the low observed rate of heart disease and
longer-than-average lifespans of people with this mutation as evidence
for it being beneficial, not its chemistry. Of course its chemistry is
also important, it avoids one saying stuff like "it could make people
more intelligent thus making them eat right" when there is actual
evidence that's not how it works at all.

(also, the chemistry is very relevant to the OP's question of whether
the mutation is "regressive" or not).

[hersheyh]

On Friday, June 14, 2013 3:54:46 AM UTC-4, jillery wrote:
> On Thu, 13 Jun 2013 23:08:37 -0700 (PDT), someone
>
> <glenn....@googlemail.com> wrote:
>
>
>
> >On http://en.wikipedia.org/wiki/Chimpanzee_genome_project it says:
>
> >
>
> >----------
>
> >Analysis of the genome was published in Nature on September 1, 2005,
>
> >in an article produced by the Chimpanzee Sequencing and Analysis
>
> >Consortium, a group of scientists which is supported in part by
>
> >the National Human Genome Research Institute, one of the National
>
> >Institutes of Health. The article marked the completion of the
>
> >draft genome sequence[2]. A database[3] now exists containing
>
> >the genetic differences between human and chimpanzee genes, with
>
> >about thirty-five million single-nucleotide changes, five million insertion/deletion events, and various chromosomal rearrangements.
>
> >Gene duplications account for most of the sequence differences
>
> >between humans and chimps. Single-base-pair substitutions account
>
> >for about half as much genetic change as does gene duplication.
>
> >----------
>
> >
>
> >Perhaps I'm misunderstanding the data here, but if we imagine that
>
> >the chimp and the human are equidistant in terms of mutations from
>
> >the common ancestor, and ignore that the chimp and the human may
>
> >have undergone the same mutations (so that there would be extra
>
> >mutations which don't show up as differences), that would still
>
> >seem to be 17.5 million single nucleotide changes that had spread
>
> >through the population and 2.5 million insertion/deletion events.
>
>
>
>
>
> Keep in mind there are approx. 3 billion (10^9) base pairs in these
>
> genomes. So even 17.5 million SNPs represents only 0.6% of the entire
>
> genome. So yes, it's likely that some SNPs are duplicates, but almost
>
> all of them are distinct locations, and the numbers reported fall
>
> within the error bars.
>
>
>
>
>
> >Does this not sound a lot for around 5 million years separation?
>
>
>
>
>
> It might sound like a lot, but again, there are lots of zeroes
>
> floating around. 5 million years is 250 thousand 20-year generations.
>
> That reduces to a mutation rate of 2.4 x 10^-8 per base pair per
>
> generation

*per haploid genome*. But


>This is comparable in magnitude to the rate determined by
>
> comparing genomes across human generations, of 1.1 x 10^-8, which I
>
> found here:
>
>
>
> http://johnhawks.net/node/16570
>
>
>
>
>
> So the difference between the two methods is about double.

Accounted for by your not including the fact that humans are diploid and receive new (and different from each other) neutral mutations from both parents.

> Does that
>
> count as "a lot" to you? I'm no expert, but I suspect the difference
>
> is from the different assumptions being made.

Another way to think about it is that each living human (and chimp or human ancestor) has a haploid genome of 3.3 x 10^9 base pairs. The mutation rate is about 1.1 x 10^-8 per base pair per generation. Nearly all of those mutations will be selectively neutral. That means that each of your ancestors passed on about 33 *new* mutations in his/her haploid contribution to his/her progeny. Thus each new individual gets 66 *new* mutations from both parents. These new mutations accumulate *each and every generation* in the lineages that survive. The new progeny generation will pass on 33 of these 'parental' new mutations to the next generation (and will get 33 from the other parent), so the next generation gets 66 of these 'second-generation' mutations. And so on. Because these mutations represent only a small fraction of the genome as yet, we can ignore duplications. Because we are only looking at the successful lineages that led to you, we can ignore the fact that some of your "new" mutations will also have

become fixed in the population as a whole and will be contributed by both parents (such fixed 'new' mutations will still represent differences between you and the modern chimp). There have been a minimum of 250 X10^3 human generations since divergence from chimps (it is likely that early hominids had a shorter generation time than modern humans, but it is also possible that the observed mutation rate is less dependent on generation time than absolute time, and it is likely that divergence was more distant in the past than 5 million years).

IOW, the observations are closer to the predicted than you think. And, of course, mutations that are *selectively* useful accumulate much faster than those that are neutral.

You, unlike many other possible individuals, are the product of an unbroken string of accumulating 'new' hominid-line mutations that typically did not also occur at that site in the chimp lineage, you should have accumulated about 16.5 x 10^6 such new mutations. But so should a modern chimp. Ergo roughly 33 x 10^6 *differences* between modern human and modern chimp genomes.

[hersheyh]

On Friday, June 14, 2013 10:06:43 AM UTC-4, John Harshman wrote:
> On 6/14/13 5:42 AM, someone wrote:
>
> > On Friday, 14 June 2013 12:25:15 UTC+1, alias Ernest Major wrote:
>
> >> On 14/06/2013 11:17, someone wrote:
>
> >>
>
> >>> On Friday, 14 June 2013 07:43:24 UTC+1, Mitchell Coffey wrote:

>
> >>>> On 6/14/2013 2:08 AM, someone wrote:
>
> >>>>> On http://en.wikipedia.org/wiki/Chimpanzee_genome_project it says:
>
> >>>>> ----------
>
> >>>>> Analysis of the genome was published in Nature on September 1, 2005,
>
> >>>>> in an article produced by the Chimpanzee Sequencing and Analysis
>
> >>>>> Consortium, a group of scientists which is supported in part by
>
> >>>>> the National Human Genome Research Institute, one of the National
>
> >>>>> Institutes of Health. The article marked the completion of the
>
> >>>>> draft genome sequence[2]. A database[3] now exists containing
>
> >>>>> the genetic differences between human and chimpanzee genes, with
>
> >>>>> about thirty-five million single-nucleotide changes, five million insertion/deletion events, and various chromosomal rearrangements.
>
> >>>>> Gene duplications account for most of the sequence differences
>
> >>>>> between humans and chimps. Single-base-pair substitutions account
>
> >>>>> for about half as much genetic change as does gene duplication
>

> >>>>> ----------
>
> >>
>
> >>>>> Perhaps I'm misunderstanding the data here, but if we imagine that
>
> >>>>> the chimp and the human are equidistant in terms of mutations from
>
> >>>>> the common ancestor, and ignore that the chimp and the human may
>
> >>>>> have undergone the same mutations (so that there would be extra
>
> >>>>> mutations which don't show up as differences), that would still
>
> >>>>> seem to be 17.5 million single nucleotide changes that had spread
>
> >>>>> through the population and 2.5 million insertion/deletion events.
>
> >>>>>
>

> >>>>> Does this not sound a lot for around 5 million years separation?
>
> >>
>

> >>>> No, why would it be?
>
> >>
>
> >>> How many naturally occurring mutations that are spreading through the human
>
> >>> population are thought to have developed in the last 2000 years? I had assumed
>
> >>> it was less than 100, and so the figures seemed surprising to me,since they
>
> >>> seem to suggest that 3-4 new single nucleotide mutations are likely to spread
>
> >>> throughout the whole population on average every year, and a new insertion
>
> >>> or deletion event every year is likely to spread throughout the whole
>
> >>> population. If the population has never been so big then the rate these
>
> >>> mutations which would go on to spread throughout humanity (assuming its
>
> >>> continued existence) would be being created could be expected to be higher
>
> >>> than ever.
>
> >>
>
> >>> [snip]
>
> >>
>
> >> In a population of constant size the rate of fixation is independent of
>
> >> the size of the population. (Mutations are fixed faster in a smaller
>
> >> population, but more mutations occur in a larger population, and these
>
> >> two factor cancel.) In a shrinking population the rate of fixation is
>
> >> greater than for a constant population (the extreme case is the founder
>
> >> effect). In a growing population the rate of fixation is lower, as there
>
> >> hasn't been time for recent mutations to be fixed, and the older
>
> >> mutations represent the smaller pool originating in a smaller population.
>
> >>
>
> >> Try restating "3-4 new single nucleotide mutations are likely to spread
>
> >> throughout the whole population on average every year" as "3-4 rare
>
> >> ancestral alleles go extinct every year".
>
> >>
>
> >
>
> > I'm not sure what you mean by the rate of fixation.
>
>
>
> Fixation is the technical term for "spread through the whole
>
> population", i.e. "reach a frequency of 100%".

To nit pick, given a mutation rate of 10^-8 and a population 6x10^9, a more realistic
way of saying it is a frequency that is effectively close to 100%. In practice, of
course, since we only can analyze a very small sample of the 6x10^9 people, 100%
of the sample analyzed is close enough. ;-)

[hersheyh]

On Friday, June 14, 2013 10:55:06 AM UTC-4, alias Ernest Major wrote:

> > Using the 100% definition, I don't think any human alleles
>
> > are fixed.
>
> >
>
> I would expect that the base pairs of some stop codons are fixed, in
>
> that a mutation that allows translation to proceed further is liable to
>
> produce a non-working protein and mammalian cell regulation is finicky
>
> about gene dosage. But in a sufficiently large population fixation
>
> becomes a less clear concept.
>
Actually, suppressor tRNAs show that not to be the case. They work because
further translation often stops rather quickly and the added amino acids do
not greatly affect function. But there certainly are some amino acids which
are crucial to specific (and vital) protein function and cannot be changed to
a different function, but because of redundancy in the code, there are often
mutations that do not change the aa at that site.
>
> For rarer genetic events like insertions, deletions, inversions and
>
> translocations is sufficiently large population is bigger.
>
>
>
> --
>
> alias Ernest Major

[jillery]

Thank you for pointing out the missing factor in my math. This helps
to affirm my point that the measured number of mutations across
generations is consistent with the measured number of mutations across
species, and that neither number is all that surprising when one
considers the total number of base pairs in a human genome.