Turing Complete Protein Switches

[John Clark]

In the September 25 2020 issue of the journal Science researchers report on the invention of a sequence of switches made entirely of protein that can perform AND OR and NOT Boolean logical operations, and thus is Turing Complete, they call it Co-LOCKR.  And they were able to put this simple computer into a T-Cell antibody, and so they could activate the T-Cell only when specific conditions are met.  

Designed protein logic to target cells with precise combinations of surface antigens

By examining the antigens on the surface of a specific type of cancer cell you can distinguish cancer cells from healthy normal cells, but it's more complex than just looking for one specific antigen. However with Co-LOCKR a T-Cell could be programmed for example, to only attack cells that have antigens W OR X  AND NOT both on their surface, AND antigen Y, AND NOT antigen Z. That way the T cell would attack cancerous cells but leave normal healthy cells alone. This is almost starting to sound a little like a simplified version of one of Drexler's Nanomachines. 

 John K Clark

[Lawrence Crowell]

It makes sense. The phosphorylation of a protein changes its shape. We can think of these different conformal shapes as different logical conditions or states.

LC

[John Clark]

On Sun, Oct 4, 2020 at 12:03 PM Lawrence Crowell <goldenfield...@gmail.com> wrote:

> It makes sense. The phosphorylation of a protein changes its shape.

The linear amino acid sequence that makes up the protein changes the way it folds up even more, from a 1-D line into a complex 3-D shape. I think predicting what linear sequence of amino acids would be needed to fold into a given 3-D shape will be the first and one of the most important tasks a Quantum Computer will work on once they become large enough to become practical.

John K Clark 

 

 

[Philip Thrift]

pdf: https://www.bakerlab.org/wp-content/uploads/2020/08/Lajoie-coLOCKR2020.pdf

via

Baker Lab, Institute for Protein Design

University of Washington, Seattle.

https://www.bakerlab.org/

[Lawrence Crowell]

Quantum computers, or processors, will make more inroads into things. They have a possible big role in understanding quantum black holes and quantum complexity. Any NP problem can be worked faster, at least in principle, with a quantum computer. In working on quantum complexity I see how this problem of a 1-dim chain filling space in a complex geometry or topology has a possible bearing on the Hodge conjecture. The use of epsilon balls and a regularization scheme in quantum complexity may play some role here.

LC

[Brent Meeker]

On 10/4/2020 12:52 PM, Lawrence Crowell wrote:

Quantum computers, or processors, will make more inroads into things. They have a possible big role in understanding quantum black holes and quantum complexity. Any NP problem can be worked faster, at least in principle, with a quantum computer.

I don't think there's any proof of that.  Given any quantum computer algorithm, it is possible that there is an equally fast classical algorithm...at least that's my understanding of the state of theory.

Brent

In working on quantum complexity I see how this problem of a 1-dim chain filling space in a complex geometry or topology has a possible bearing on the Hodge conjecture. The use of epsilon balls and a regularization scheme in quantum complexity may play some role here.

LC

On Sunday, October 4, 2020 at 1:54:45 PM UTC-5 johnk...@gmail.com wrote:

On Sun, Oct 4, 2020 at 12:03 PM Lawrence Crowell <goldenfield...@gmail.com> wrote:

> It makes sense. The phosphorylation of a protein changes its shape.

The linear amino acid sequence that makes up the protein changes the way it folds up even more, from a 1-D line into a complex 3-D shape. I think predicting what linear sequence of amino acids would be needed to fold into a given 3-D shape will be the first and one of the most important tasks a Quantum Computer will work on once they become large enough to become practical.

John K Clark 

 

 

--
You received this message because you are subscribed to the Google Groups "Everything List" group.
To unsubscribe from this group and stop receiving emails from it, send an email to everything-li...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/everything-list/40a97d29-ff4d-42de-8d5f-a449d3c5fdben%40googlegroups.com.

[Lawrence Crowell]

On Sunday, October 4, 2020 at 6:04:32 PM UTC-5 Brent wrote:

On 10/4/2020 12:52 PM, Lawrence Crowell wrote:

Quantum computers, or processors, will make more inroads into things. They have a possible big role in understanding quantum black holes and quantum complexity. Any NP problem can be worked faster, at least in principle, with a quantum computer.

I don't think there's any proof of that.  Given any quantum computer algorithm, it is possible that there is an equally fast classical algorithm...at least that's my understanding of the state of theory.

Brent

Quantum computers have a polynomial speed up of NP problems. There is though no general rule for this, and some NP problems are not much faster. Quantum computers are bounded quantum polynomial which have some space/time advantages over classical computers.

LC

[John Clark]

On Sun, Oct 4, 2020  'Brent Meeker' via Everything List <everyth...@googlegroups.com> wrote:

On 10/4/2020 12:52 PM, Lawrence Crowell wrote:

>> Quantum computers, or processors, will make more inroads into things. They have a possible big role in understanding quantum black holes and quantum complexity. Any NP problem can be worked faster, at least in principle, with a quantum computer.

> I don't think there's any proof of that.  Given any quantum computer algorithm, it is possible that there is an equally fast classical algorithm

It's true that although a quantum algorithm has been found that can factor numbers efficiently there is no proof a classical algorithm cannot be discovered that would do the same thing, in fact it has never been proven that P≠NP, although nearly all mathematicians believe that is the case.  However it has been proven that a recently discovered exotic class of problems can be solved In polynomial time but even if it turns out to everybody's surprise that P=NP and a classical algorithm is found to make use of that fact a classical computer could never do as well solving them as a quantum computer. It's so new that nobody is yet quite sure if this exotic class of problems is of interest in themselves or is interesting only because a conventional computer could not solve them efficiently but a quantum computer could. Although falling short of a proof it gives yet more ammunition to those who believe a quantum computer can solve more familiar practical problems faster than a classical computer ever will be able to.

Oracle Separation of BQP and PH

I think the killer application for a quantum computer will be simulating quantum systems. 

John K Clark 

[Philip Thrift]

On Monday, October 5, 2020 at 1:55:20 PM UTC-5 johnk...@gmail.com wrote:

I think the killer application for a quantum computer will be simulating quantum systems. 

John K Clark 

Why shouldn't simulations of quantum systems on (massive CPU/GPU parallel) computers be just as good?

@philipthrift 

[John Clark]

On Mon, Oct 5, 2020 at 3:14 PM Philip Thrift <cloud...@gmail.com> wrote:

>> I think the killer application for a quantum computer will be simulating quantum systems. 

> Why shouldn't simulations of quantum systems on (massive CPU/GPU parallel) computers be just as good?

Because regardless of how big your classical computer is if you're running any known classical algorithm on it every time you Increase the number of interacting particles by 1 in a quantum system (N) you increase the number of computations required to simulate it exponentially, but with a quantum computer the increase is only geometrical. Basically it's the difference between 2^N and N^2, one increases MUCH faster than the other.  So even a supercomputer, if it's classical, can only simulate very simple quantum systems exactly, very soon you need to make all sorts of simplifying approximations, and not long after that you can't even make approximations that are worth a damn. A classical computer can't even figure out the freezing point of water starting from Schrodinger's Equation because it's just too complicated to calculate, we need to do experiments if we want to know it. And that's a molecule involving only 3 atoms and 10 electrons.

John K Clark

[Lawrence Crowell]

On Monday, October 5, 2020 at 1:55:20 PM UTC-5 johnk...@gmail.com wrote:

The physical idea is that a quantum computer is a faster is that in principle if we had quantum brains it would really be exponentially faster. However, the result of a quantum computation can only be be manifested if the entanglements are decoded by a classical signal. This "undoes" the exponential speed up. However, for teleporting a state with a Bell pair the classical part has half the information. Then in principle, for quantum computing there is a speed up that is some fraction of what occurs with a classical computer. The actual speed up is dependent on the algorithm as well. 

This paper on BQP and PH made a related point in that BQP has the need for fewer oracle inputs, which is the same as saying user inputs. This means for a range of problems quantum computing will have an economy of time or scale. 

Certainly right away quantum computing will be mostly used for modelling systems, in particular quantum systems. Quantum computing has analogues with black holes as well. The complexity of computing has analogues with quantum complexity of black holes. We may then have laboratory-like forms or simulations of black holes with quantum computers. In fact I think this can happen with a certain optical process with atoms that can make an atom quantum computing.

LC

[Philip Thrift]

Quantum computers may end up being not-more-advantageous for "quantum modeling" than the current trend in using DLNs (deep leaning nets).

e.g. https://arxiv.org/abs/1909.02487

@philipthrift

[Bruno Marchal]

On 4 Oct 2020, at 14:07, John Clark <johnk...@gmail.com> wrote:

In the September 25 2020 issue of the journal Science researchers report on the invention of a sequence of switches made entirely of protein that can perform AND OR and NOT Boolean logical operations, and thus is Turing Complete, they call it Co-LOCKR.  And they were able to put this simple computer into a T-Cell antibody, and so they could activate the T-Cell only when specific conditions are met.  

I can accept this, in a very loose sense of “Turing Complete”. But normally “Turing complete” is an attribute of a theory, and “Turing universal” is for a computer/machine. The NOR is not Turing-complete, but the combination of them does provide a Turing universal machinery. It might mean that a composition of such protein could constitute un computer. I doubt that one unique protein can be Turing universal, although (and I did prove this) a sufficiently long DNA + some cytoplasme, including some protein, is Turing Universal (and it makes any theory capable of desiring this Turing-complete).

I mention this only because when we apply the notion of Turing universality and Turing completeness in metaphysics, at some point those nuances get important to keep in Mind, and that is why now I distinguish in my posts the notion of universal machinery (like the RE collection of all Turing machine (all set of quadruplets)) and of universal machine (one precise set of quadruplets). 

When we have a universal machine or function phi_u, we have a universal machinery: phi_u(I, x), i = 0, 1, 2, …

And Turing’s work gives the reverse: when we have a universal machinery, we get a universal machine among it.

Those things are equivalent with respect of computability, but they are used at different level, and in metaphysics, we need to postulate a universal machinery, before a universal machine.

Designed protein logic to target cells with precise combinations of surface antigens

By examining the antigens on the surface of a specific type of cancer cell you can distinguish cancer cells from healthy normal cells, but it's more complex than just looking for one specific antigen. However with Co-LOCKR a T-Cell could be programmed for example, to only attack cells that have antigens W OR X  AND NOT both on their surface, AND antigen Y, AND NOT antigen Z. That way the T cell would attack cancerous cells but leave normal healthy cells alone. This is almost starting to sound a little like a simplified version of one of Drexler's Nanomachines. 

OK. Looks quite interesting. It could help also in the problem of the origin of life, although here it is the RNA which would need to be “Turing complete” (in that loose sense of belonging to a possible universal machinery).

In principle, one protein + one DNA strands (that the protein should be able to elongated if needed) can be a genuine Turing universal machine. They always need some elongate-able tape...

Bruno

 John K Clark

--
You received this message because you are subscribed to the Google Groups "Everything List" group.
To unsubscribe from this group and stop receiving emails from it, send an email to everything-li...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/everything-list/CAJPayv3Au%2BZFSRc%2B8xm%3Dhj%2B_f3wOff%2Bi61nLS3_1nwwPBP1tKA%40mail.gmail.com.

[Bruno Marchal]

On 5 Oct 2020, at 01:04, 'Brent Meeker' via Everything List <everyth...@googlegroups.com> wrote:

On 10/4/2020 12:52 PM, Lawrence Crowell wrote:

Quantum computers, or processors, will make more inroads into things. They have a possible big role in understanding quantum black holes and quantum complexity. Any NP problem can be worked faster, at least in principle, with a quantum computer.

I don't think there's any proof of that.  Given any quantum computer algorithm, it is possible that there is an equally fast classical algorithm...at least that's my understanding of the state of theory.

P ≠ NP is like Riemann Hypothesis. We don’t have (yet) a proof, but very few people believe that it could be false.

But also, even if P = NP, it might be that in practice the polynomial will be so complex that a quantum computer will still run faster than any reasonable classical computer. Now, I sincerely doubt that P could be equal to NP, despite the vexing difficulty we met when trying to prove P ≠ NP.

Bruno

Brent

In working on quantum complexity I see how this problem of a 1-dim chain filling space in a complex geometry or topology has a possible bearing on the Hodge conjecture. The use of epsilon balls and a regularization scheme in quantum complexity may play some role here.

LC

On Sunday, October 4, 2020 at 1:54:45 PM UTC-5 johnk...@gmail.com wrote:

On Sun, Oct 4, 2020 at 12:03 PM Lawrence Crowell <goldenfield...@gmail.com> wrote:

> It makes sense. The phosphorylation of a protein changes its shape.

The linear amino acid sequence that makes up the protein changes the way it folds up even more, from a 1-D line into a complex 3-D shape. I think predicting what linear sequence of amino acids would be needed to fold into a given 3-D shape will be the first and one of the most important tasks a Quantum Computer will work on once they become large enough to become practical.

John K Clark 

 

 

--
You received this message because you are subscribed to the Google Groups "Everything List" group.
To unsubscribe from this group and stop receiving emails from it, send an email to everything-li...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/everything-list/40a97d29-ff4d-42de-8d5f-a449d3c5fdben%40googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Everything List" group.
To unsubscribe from this group and stop receiving emails from it, send an email to everything-li...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/everything-list/9443b9af-3792-ca2f-bd04-f51653d45612%40verizon.net.

[Bruno Marchal]

On 4 Oct 2020, at 20:56, Philip Thrift <cloud...@gmail.com> wrote:

pdf: https://www.bakerlab.org/wp-content/uploads/2020/08/Lajoie-coLOCKR2020.pdf

Thanks. This confirms somehow what I said. 

Impressive work. No doubt about that. 

Bruno

--
You received this message because you are subscribed to the Google Groups "Everything List" group.
To unsubscribe from this group and stop receiving emails from it, send an email to everything-li...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/everything-list/8bfc52b3-520d-4602-ad85-54c5af691976n%40googlegroups.com.

[Bruno Marchal]

It is good enough! …in theory. 

You can simulate a quantum computer classically, but to factorise a number with more than one or two hundreds of digits, the observable universe is not big enough to contain the massive parallel classical computers. 

This means that even the quasi-exact simulation of a proton requires something bigger than a (single) physical universe (if that notion could make sense).

Now, as Deutsch has shown, any quantum system can be emulated by a quantum computer in polynomial time.

Bruno

@philipthrift 

--
You received this message because you are subscribed to the Google Groups "Everything List" group.
To unsubscribe from this group and stop receiving emails from it, send an email to everything-li...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/everything-list/d737022e-8535-4d75-afca-86466aeb1fd6n%40googlegroups.com.

[Brent Meeker]

On 10/6/2020 4:35 AM, Bruno Marchal wrote:
>
> OK. Looks quite interesting. It could help also in the problem of the
> origin of life, although here it is the RNA which would need to be
> “Turing complete” (in that loose sense of belonging to a possible
> universal machinery).

In the origin of life it must be the system, including some aspects of
the environment that constitute Turing complete or near Turing complete
machinery.  Nick Lane is an advocate of metabolism first has written a
nice popular book on the origin of life "The Vital Question" in which
the ATP cycle precedes RNA.

Brent