folding proteins? - a newbie question about nanotechnology
Burak
i cant seem to understand why creating proteins that always fold into
particular forms is so important? I mean if you already have the
technology to position a single atom accurately (i saw the ibm logo
with the xenon atoms), why do you need these folding proteins? Is it
related to creating certain types of bonds for different purposes? is
it sth intended to smooth, or just expedite the manufacturing process
by using larger and standard building blocks? or is it intended just
for manufacturing certain types of nanomachines such as the ones with
actual moving parts like gears, hinges, etc.?
I am just an enthusiast about nanotechnology and i would really
appreciate if someone could answer my question. The articles i have
seen so far just does not seem to indicate the difference, or i am
just missing some point somewhere.
Thanks,
Burak
Gordon D. Pusch
bura...@hotmail.com (Burak) writes:
> i cant seem to understand why creating proteins that always fold into
> particular forms is so important?
Aside from the fact that we would very much like to figure out how to make
"designer proteins" ab initio, it is because each protein strand can fold
up into a multitude of conformations, and the overwhelming majority of
these conformations will either _NOT_ have the desired physical properties or
catalytic activity, or even worse, might have the _WRONG_ catalytic activity.
(Example: Prion diseases such as KJD or BSE (AKA "Mad Cow Disease") are
caused by a protein that folded up the "wrong way," which catalyzes
_other_ proteins of the same type to convert from the "right" fold
to the "wrong" fold.) Hence, proteins that fold up the "wrong way"
may not only _NOT_ work as enzymes or medicines, they may actually be
poisonous, or cause diseases, Which Would Be Bad.
> I mean if you already have the technology to position a single atom
> accurately (i saw the ibm logo with the xenon atoms),
In point of fact, we do =NOT= yet have the technology to "position single
atoms accurately." We have a _VERY_ crude technology that can _ROUGHLY_
position _CERTAIN_ species of atoms such as Xenon on _SOME_ 2-D substrates
such as silicon, at _CRYGOGENIC TEMPERATURES_ and _IN A VACUUM_.
We still have a =LONG= way to go before we will be able to position
an _ARBITRARY_ species of atom in _THREE-DIMENSIONAL SPACE_ at
_ROOM TEMPERATURE_ and in a _WATER_ environment --- which is what
is required to build up proteins "atom by atom," and have them
be folded up correctly. Until such time as "strong" nanotech is achieved,
all we currently have is "naturally evolved" nanotech: i.e., proteins and
their biological synthesis.
> why do you need these folding proteins?
To catalyze desired chemical reactions at more reasonable temperatures and
pressures than current "brute force" chemical engineering. To serve as drugs.
To act as supplements for people whose bodies do not make enough of that type
of protein, or whose bodies make a damaged form of that protein, such as in
persons suffering from cystic fibrosis or sickle-cell anemia. To create new
types and forms of "biomaterials" to repair or augment biological tissue.
And to carry out a host of other possible applications that I don't have
time to list right now --- so please consult Robert Frietas' "Nanomedicine"
or the extensive literature on "wet" nanotech, AKA "biotech."
> Is it related to creating certain types of bonds for different purposes?
Among other things. For example, that is one of the ways living cells use
proteins.
> is it sth intended to smooth, or just expedite the manufacturing process
> by using larger and standard building blocks?
That would certainly be one application, and it appears to be one of the
reasons why living cells manufacture such "standardized building blocks."
> or is it intended just for manufacturing certain types of nanomachines
> such as the ones with actual moving parts like gears, hinges, etc.?
Again, that appears to be why living cells make proteins.
> I am just an enthusiast about nanotechnology and i would really
> appreciate if someone could answer my question. The articles i have
> seen so far just does not seem to indicate the difference, or i am
> just missing some point somewhere.
You are missing the difference between "dry" and "wet" nanotech, as well
as the fact that we do not yet _have_ functioning "dry" nanotech, while
in some limited respects, we _do_ have "wet" nanotech, i.e., biotech.
-- Gordon D. Pusch
perl -e '$_ = "gdpusch\@NO.xnet.SPAM.com\n"; s/NO\.//; s/SPAM\.//; print;'
erincss
Hello. I have a question for Gordon and everyone else here. I have been reading
about the development of quantum computer technologies, and how this applies to
Molecular Manufacturing, and it seems that many of the people in the Q
Computation field believe that sufficiently advanced quantum computers will
once and for all be able to solve the Protein Folding puzzle, and will allow us
to design arbitrary proteins that fold up into three dimensional structures.
What is the realism of this? I would like all of your thoughts.
Summary: Does possessing a true quantum computer automatically mean protein
folding problems are solved?
Thank you and take care :)
Gordon D. Pusch
eri...@aol.com (erincss) writes:
> I have been reading about the development of quantum computer technologies,
> and how this applies to Molecular Manufacturing, and it seems that many
> of the people in the Q Computation field believe that sufficiently
> advanced quantum computers will once and for all be able to solve the
> Protein Folding puzzle, and will allow us to design arbitrary proteins
> that fold up into three dimensional structures.
>
> What is the realism of this? I would like all of your thoughts.
My own opinion is that such claims are vastly overblown, in part because of
the false "folk theorem" that quantum computers are "more powerful" than
classical computers in some fundamental way. On the contrary, an N-qubit
quantum computer can be _EXACTLY_ simulated by a 2^N bit classical computer,
so for bounded resources, the so-called "quantum speedup" begins to look
more like an exotic form of "time vs. memory" tradeoff than any sort of
fundamental advance in the nature of what is considered "computable."
Another over-hyped aspect of quantum computing is the belief by some that
quantum computers can somehow compute "all the branches of a program in
parallel" and then preferentially "collapse" the computation into the
"correct" branch. This false belief usually stems (IMO) from either an
incorrect physical understanding of the nature of quantum superposition,
or a too-literal reading of the "Many Worlds Interpretation" of quantum
mechanics. On the contrary, it is impossible for a quantum system to
"choose" what state it collapses into; rather, if it is in a quantum
superposition, it will collapse into _any_ of its possible outcomes
_at random_, with some probability in principle determined by the state
preparation. Hence, to "prepare" a quantum computer so it "collapses"
to the "correct" state would require that one already know the "correct"
outcome _IN ADVANCE_, which rather obviates the point of computing anything!
Likewise, if the various branches of the computation are distinguishable
(which they _MUST_ be if one is to be able to tell their results apart!),
then they _cannot_ be superposed in the fashion that such authors believe,
since distinguishable outcomes _CANNOT_ interfere with each other;
the classic example of this is the so-called "Welcher Weg" ("which way")
variant of the "Two slit" experiment, wherein determining which path the
quantum particle took destroys the quantum superposition and reduces the
two-slit pattern to two overlapping one-slit patterns.
> Summary: Does possessing a true quantum computer automatically mean
> protein folding problems are solved?
Personally, I very much doubt it.
John S. Novak, III
In article <c15c1...@enews3.newsguy.com>, Burak wrote:
> i cant seem to understand why creating proteins that always fold into
> particular forms is so important? I mean if you already have the
> technology to position a single atom accurately (i saw the ibm logo
> with the xenon atoms), why do you need these folding proteins? Is it
> related to creating certain types of bonds for different purposes? is
> it sth intended to smooth, or just expedite the manufacturing process
> by using larger and standard building blocks? or is it intended just
> for manufacturing certain types of nanomachines such as the ones with
> actual moving parts like gears, hinges, etc.?
ability to place large numbers of atoms precisely right now, we might
not ever. It is an open question of engineering, right now.
Whie I don't believe this to be the case, protein folding might be the
closest we can get to this ideal, and therefore directly merit study
for that reason. Less directly, we can certainly gain nisight nito
molecular positioning problems by looking at a class of systems which
already does just that. Finally, if there are multiple paths to
precise atomic placement (say, Drexlerian and protein folding, to keep
things simple) then there are very likely to be situations where one
is a better choice than the other. This may be for purely technical
reasons, or it may be for economic reasons.
Second, proteins are all around us. Proteins are the mechanism os
human life. Anything we learn about protein folding will return
extreme practical dividends in the health and life sciences. If we
could design folded proteins (or even analyze them!) like we're soon
going to be able to sequence genomes, many human health problems would
be much more tractable than they are now.
--
John S. Novak, III j...@cegt201.bradley.edu
The Humblest Man on the Net
John S. Novak, III
I'm not an expert in either field, but I know a little bit about both.
I can't see any reason to assume that once quantum computers are a
reality, that protein folding problems are automatically solved.
In fact, once quantum computers are a reality, I don't see a reason to
believe that very many areas of computer science are automatically
solved or altered, except for a few security and communication areas.
Quantum computers are not magic, and it is fiendishly difficult to
come up with quantum algorithms that really improve on classical ones.
There is a marked tendency for enthusiasts of a particular field
(protein folding, molecular computing, Drexlerian dry nanotech,
quantum computation, genetic algorithms, etc) to treat the broad
approaches as Holy Grails that will solve all the world's problems as
soon as we just figure out this one detail.
Editorial mode, on:
Trust me, folks, and trust every single practicing engineer outside of
academia (and most of the ones in academia): It ain't that simple.
Reality is much grittier than the tech-pundits make it out to be.
Sample anecdote: I decided a few years ago that I wanted to be in a
better position to do research on nanotechnology. I decided the
easiest way to do that was to *go get a PhD in computer science.* And
that the easiest way to do that would be to keep working as an
electrical engineer with an MSEE while simultaneously breaking my back
getting an MSCS in my copious free time.
After about four years, I have clawed myself into enough knowledge,
skill, and good grace with the faculty members, to start doing little
invetigations into protein folding, and am perhaps six months away
from the MSCS (although I may stetch it out a little and take more
classes if it suits me.)
This is hard stuff, folks.
(The good side is, I'm so busy I don't have time to spend my salary,
so I can store it away for the day I quit and go to school full time
for the PhD.)
No one approach is going to solve everything. Every new approach may
make some other problem a little easier, but each step is the result
of lots and lots of hard work, and everything in research is
customized-- by definition.
I think this is why some people get frustrated by what they see as the
unrelenting hype of the media-- it makes everything look easy and just
around the corner.
Editorial mode, off.
Bob
Other people have addressed the over-optimism about the development of
Q.C. I'd like to briefly address the other part of that. I am not
convinced that protein folding can be solved computationally at all.
To do so requires specifying all of the needed information -- eg, the
strengths of all the possible interactions. But that is a very long
list. Further, the precise value of the strength of, say, a hydrogen
bond depends on the precise details, including solvent interactions.
Protein stability is quite low, and the energy of the folded state is
a subtle choice among many similar states.
Of course, we will learn more and more, and will solve favorable
cases, etc. But I think that any claim that the protein folding
problem will be easily solved simply by throwing more computational
power after it is overblown.
bob
John S. Novak, III
In article <c1bo1...@enews1.newsguy.com>, Bob wrote:
> Other people have addressed the over-optimism about the development of
> Q.C. I'd like to briefly address the other part of that. I am not
> convinced that protein folding can be solved computationally at all.
> To do so requires specifying all of the needed information -- eg, the
> strengths of all the possible interactions. But that is a very long
> list. Further, the precise value of the strength of, say, a hydrogen
> bond depends on the precise details, including solvent interactions.
> Protein stability is quite low, and the energy of the folded state is
> a subtle choice among many similar states.
associated with protein folding is non-computable. I'm not sure if
it's the analytic claim (how does this protein fold?) or the synthetic
claim (how do we get a protein to fold thus?) but in either case, I'm
not sure if I believe the claim.
I certainly don't believe that the analytic claim is true-- after all,
nature makes these computations, if you will, on a regular basis.
It's just physics.
I might be persuaded to believe the synthetic claim, but it would take
a better argument than "It's tough." I happen to know that solvent
interactions are a standard part of protein folding analyses today:
They're certainly not being left out of the picture; I have a stack of
papers and references to solvent accessible surface area calculations
on my desk right now.
> Of course, we will learn more and more, and will solve favorable
> cases, etc. But I think that any claim that the protein folding
> problem will be easily solved simply by throwing more computational
> power after it is overblown.
Reputable people, I think, are claiming that massive computational
power is a necessary but not sufficient condition. But I do agree
with your statement.
Bob
On 23 Feb 2004 06:05:25 GMT, j...@panix.com (John S. Novak, III) wrote:
>
>In article <c1bo1...@enews1.newsguy.com>, Bob wrote:
>
>> Other people have addressed the over-optimism about the development of
>> Q.C. I'd like to briefly address the other part of that. I am not
>> convinced that protein folding can be solved computationally at all.
>> To do so requires specifying all of the needed information -- eg, the
>> strengths of all the possible interactions. But that is a very long
>> list. Further, the precise value of the strength of, say, a hydrogen
>> bond depends on the precise details, including solvent interactions.
>> Protein stability is quite low, and the energy of the folded state is
>> a subtle choice among many similar states.
>
>Somewhere in here, I think you are making a claim that something
>associated with protein folding is non-computable. I'm not sure if
>it's the analytic claim (how does this protein fold?) or the synthetic
>claim (how do we get a protein to fold thus?) but in either case, I'm
>not sure if I believe the claim.
>
Yes, I am claiming that one may well not be able to compute how a
protein will fold -- in the general case. That is not a matter of
principle, but of having the data necessary to do the computation. One
would need, for example, the energy of each H-bond (not of "average
H-bonds) -- under the relevant conditions. Again, remember that the
net energy of folding may only be a few kJ/mol (on the order of a
single H-bond) -- and the difference between right and wrong folded
states may be much less. That net energy comes from summing thousands
of terms.
>I certainly don't believe that the analytic claim is true-- after all,
>nature makes these computations, if you will, on a regular basis.
>It's just physics.
>
Nature is quite remarkable. But it has had lots of time to work on the
problem. And it has a powerful tool for working it out (natural
selection), which means that it need not know the data. Yet it is
still estimated that something like 1/3 of the newly made proteins in
humans are rapidly degraded, with some fraction of those presumably
due to having not folded properly.
>I might be persuaded to believe the synthetic claim, but it would take
>a better argument than "It's tough."
How about "It's _very_ tough"?
>I happen to know that solvent
>interactions are a standard part of protein folding analyses today:
>They're certainly not being left out of the picture; I have a stack of
>papers and references to solvent accessible surface area calculations
>on my desk right now.
I think you just made my point. :-) They try to take into account
solvent interactions, and they do not succeed in solving the problem.
In fact, the statement presumes that one clearly understands what the
appropriate solvent is.
Working on a problem requires some optimism that what one is doing is
going to be useful. (Of course, that optimism also is needed to get
grants funded. But I am not trying to be cynical here.) The work
should continue. Presumably we will continually make progress. The
context of the original question here was whether a (particular) magic
solution was imminent.
bob
John S. Novak, III
In article <c1ekj...@enews1.newsguy.com>, Bob wrote:
>>Somewhere in here, I think you are making a claim that something
>>associated with protein folding is non-computable. I'm not sure if
>>it's the analytic claim (how does this protein fold?) or the synthetic
>>claim (how do we get a protein to fold thus?) but in either case, I'm
>>not sure if I believe the claim.
> Yes, I am claiming that one may well not be able to compute how a
> protein will fold -- in the general case. That is not a matter of
> principle, but of having the data necessary to do the computation. One
> would need, for example, the energy of each H-bond (not of "average
> H-bonds) -- under the relevant conditions. Again, remember that the
> net energy of folding may only be a few kJ/mol (on the order of a
> single H-bond) -- and the difference between right and wrong folded
> states may be much less. That net energy comes from summing thousands
> of terms.
Well, sure.
That's why we don't use pencil and paper to do these calculations.
My electromagnetic simulations at work are the result of millions of
operations, but that makes the problems neither more nor less
computable than protein folding.
>>I certainly don't believe that the analytic claim is true-- after all,
>>nature makes these computations, if you will, on a regular basis.
>>It's just physics.
> Nature is quite remarkable. But it has had lots of time to work on the
> problem. And it has a powerful tool for working it out (natural
> selection), which means that it need not know the data. Yet it is
> still estimated that something like 1/3 of the newly made proteins in
> humans are rapidly degraded, with some fraction of those presumably
> due to having not folded properly.
You don't seem to understand.
I'm not talking about the synthesis problem of designing a protein in
advance. I'm very specifically talking about the analysis problem of
taking the sequence for a protein and predicting how it will fold.
This is very clearly a computable problem, because the proteins
themselves are computing the result millions of times over every
second. At no point do the atoms of a given protein *fail* to take
some shape. The answer is the answer.
>>I might be persuaded to believe the synthetic claim, but it would take
>>a better argument than "It's tough."
> How about "It's _very_ tough"?
No.
Computability is a term of art in computer science, with a rigourous
meaning. Appeals to conventional measures of difficulty do not make
an argument to non-computability.
Especially not when we're talking about a known physical process.
>>I happen to know that solvent
>>interactions are a standard part of protein folding analyses today:
>>They're certainly not being left out of the picture; I have a stack of
>>papers and references to solvent accessible surface area calculations
>>on my desk right now.
> I think you just made my point. :-) They try to take into account
> solvent interactions, and they do not succeed in solving the problem.
> In fact, the statement presumes that one clearly understands what the
> appropriate solvent is.
No.
Your answer devolves to a fancy form of, "People have been working on
it and haven't gotten it. It's not computable."
This is, to be blunt, a false argument.
> Working on a problem requires some optimism that what one is doing is
> going to be useful.
Many graduate students will disagree with this statement.
Gordon D. Pusch
Bob <bbr...@uclink4.berkeley.edu> writes:
....And beyond even the question of whether the protein-folding problem
is "computable," or even "computable give data that we do not yet have,"
there is the issue alluded to in my own post, namely, that the conformation
Nature "chooses" may _NOT_ necessarily be the "minimum energy" conformation,
as tacitly assumed in most "protein folding" calculations. The existence
of "prion" diseases such as KJD or BSE wherein an alternate conformation
catalyzes the re-folding of the "natural" conformation of a protein into
the "prion" conformation strongly suggests that the "natural" protein
conformation used in living cells is in some cases _NOT_ the "minimum
energy" state of that protein, but rather an alternate, metastable
conformation with a higher energy instead, and is capable of spontaneously
re-folding into a lower energy state given the right "nudge." (The putative
existence of "chaperonin" proteins that are believed to "guide" the folding
of some other protein into particular desired conformation also strongly
suggests that proteins do _not_ always spontaneously fold up into the
"correct" or "minimum energy" conformation by themselves...)
Bob
On 24 Feb 2004 06:09:13 GMT, j...@panix.com (John S. Novak, III) wrote:
>
>In article <c1ekj...@enews1.newsguy.com>, Bob wrote:
>
>>>Somewhere in here, I think you are making a claim that something
>>>associated with protein folding is non-computable. I'm not sure if
>>>it's the analytic claim (how does this protein fold?) or the synthetic
>>>claim (how do we get a protein to fold thus?) but in either case, I'm
>>>not sure if I believe the claim.
>
>> Yes, I am claiming that one may well not be able to compute how a
>> protein will fold -- in the general case. That is not a matter of
>> principle, but of having the data necessary to do the computation. One
>> would need, for example, the energy of each H-bond (not of "average
>> H-bonds) -- under the relevant conditions. Again, remember that the
>> net energy of folding may only be a few kJ/mol (on the order of a
>> single H-bond) -- and the difference between right and wrong folded
>> states may be much less. That net energy comes from summing thousands
>> of terms.
>
>Well, sure.
>That's why we don't use pencil and paper to do these calculations.
>
>My electromagnetic simulations at work are the result of millions of
>operations, but that makes the problems neither more nor less
>computable than protein folding.
>
You are missing the point. The problem is not the number of
calculation steps, but the lack of data for the numbers to be
calculated. I hinted at this above, but may have assumed you
understood something about protein energies; if you are more of a
physical scientist, it may have been too subtle. Sorry.
The thousands of terms being summed total, say, 10 kJ per mole -- and
each of those terms is of the general order of, say, 1-10 kJ/mole.
[What we need to calculate is the difference in energy between one
folded state and another.] Thus it logically follows that these terms
must be known to exceptional precision. It is basically a problem of
small differences between large numbers. Energies are not usually
known to that kind of precision. Now, I stretch that point, and
suggest that it may be very difficult to ever know them that well.
Whether my stretch is correct or not is something we will find out
over the decades with experience.
(Of course, one could also note that energy states within RT or so of
each other may well co-exist. That proteins may not have a simple
single stable state is probably a good idea, but maybe is just too
complicated for the moment.)
>>>I certainly don't believe that the analytic claim is true-- after all,
>>>nature makes these computations, if you will, on a regular basis.
>>>It's just physics.
>
>> Nature is quite remarkable. But it has had lots of time to work on the
>> problem. And it has a powerful tool for working it out (natural
>> selection), which means that it need not know the data. Yet it is
>> still estimated that something like 1/3 of the newly made proteins in
>> humans are rapidly degraded, with some fraction of those presumably
>> due to having not folded properly.
>
>You don't seem to understand.
>I'm not talking about the synthesis problem of designing a protein in
>advance. I'm very specifically talking about the analysis problem of
>taking the sequence for a protein and predicting how it will fold.
so am I. But I agree that the design problem is even more difficult.
>
>This is very clearly a computable problem, because the proteins
>themselves are computing the result millions of times over every
>second.
That is cute, but it is also quite wrong if you start taking it
literally. Proteins do not _compute_ anything. (And the earth does not
compute where to go next in its orbit.) They simply do, or fold. It is
stochastic. Some fold correctly, some do not. Some of the latter may
be given a chance to re-try. Some sequences are incapable of folding
in any well-defined way at all. Of course, mother nature has
eliminated these from the gene pool. (This point has been shown by
Monte Carlo simulations of folding of random sequence chains. Most
don't fold.) There are additional complications, such as unstructured
regions (which may develop into an important structure after something
is bound), flexibility, effect of environmental conditions.
>At no point do the atoms of a given protein *fail* to take
>some shape. The answer is the answer.
well, yes. But now the Q is whether we know how -- or will be able to
know how -- to calculate that answer. And I am suggesting that we do
not (and perhaps will not) know the data needed to do the
calculations. To go back to your analogy that the proteins do the
computations... they have the data, and we do not.
(Someone -- I forget the source -- quipped... Nature does not have a
protein folding problem. We do.)
>
>>>I might be persuaded to believe the synthetic claim, but it would take
>>>a better argument than "It's tough."
>
>> How about "It's _very_ tough"?
>
>No.
>
>Computability is a term of art in computer science, with a rigourous
>meaning. Appeals to conventional measures of difficulty do not make
>an argument to non-computability.
>
>Especially not when we're talking about a known physical process.
I have said since the beginning of this discussion that the problem
can be calculated in principle. The barrier is not the principle, but
having the data needed to make the calculation.
If I ask you how long it will take for me to take "the trip", you
would reply that it is computable. Fine, but unless you know the rate
and distance, you would not be able to actually do the computation.
That is the situation with calculating protein folding.
>
>>>I happen to know that solvent
>>>interactions are a standard part of protein folding analyses today:
>>>They're certainly not being left out of the picture; I have a stack of
>>>papers and references to solvent accessible surface area calculations
>>>on my desk right now.
>
>> I think you just made my point. :-) They try to take into account
>> solvent interactions, and they do not succeed in solving the problem.
>> In fact, the statement presumes that one clearly understands what the
>> appropriate solvent is.
>
>No.
>Your answer devolves to a fancy form of, "People have been working on
>it and haven't gotten it. It's not computable."
>
>This is, to be blunt, a false argument.
I understand the point you are trying to make. But you are missing my
point -- which is based on understanding biology. I am suggesting that
it may, for all practical purposes, be impossible to determine the
needed info. Now, saying it with that conciseness makes it sound
extreme. As I said, my brief form of the statement is a stretch. But
the stretch is of quite well understood issues. (That is, my claim may
be hyperbole, but only in degree. And whether it is or not is
something only time will tell.)
Perhaps the issue is... should we just keep doing more of what we are
doing (get more computing power), or should we analyze the problem for
predictable barriers/problems?
>
>> Working on a problem requires some optimism that what one is doing is
>> going to be useful.
>
>Many graduate students will disagree with this statement.
:-). I chose my own project!
Physicists promoted a revolution in biology in the mid 20th century.
They did this by the reductionist approach, and an insistence on
quantitation (something biologists have traditionally been rather
casual about). All the basics of our understanding of molecular
genetics came from this approach. OTOH, almost everything we learned
from that approach is wrong. Not wrong in the sense it has to be
thrown out, but in the sense that it is over-simplified. So
over-simplified that we could not really understand how organisms
work, without a new revolution more on the other side, emphasizing the
complexities. I still enjoy teaching the story of the lac operon; it
is a wonderful story of discovery. But I do cringe at how wrong the
model really is -- it taken too literally. Of course, taking the step
to a simple model is important. But that is not the answer.
I think that non-biologists fail to understand the complexities and
subtleties of the protein folding problem. It is easy to think that
the old simple story that a protein simply folds into a well-defined
low energy state is the entire story. Chaperones, flexibility,
degradation of the substantial amount of improperly folded proteins --
these things surprised the biologists, and seem to not have been
grasped by those on the outside.
bob