D-Wave goes public with 81-qubit protein modeling D-Wave – whose claims to have a working quantum computer have been met with skepticism and major contracts in equal measure – has published a paper in Nature in which it demonstrates the application of quantum annealing to protein folding analysis. Protein folding is a difficult problem in the classical world, because of the …
If they actually have a working computer that could speed up protein folding calculation, I'm moving to whichever country* buys this computer first. Not because I would expect any immediate gains from this, but because I want my taxes to pay for this awesomeness.
Still a bit sceptical, though.
*Except for Murrkah or Switzerland.
"protein folding calculation"
It's not a 'calculation' really -just a series of relatively poor algorithms. Huge strides in the basic science have taken place in recent years - these have been assisted by the huge increase in processing power but it's all still far from routine - it's often (relatively) easy to fold a new protein if it has homology to a known structure but even in that 'easy' case it's often found that the optimized solution is still a poor fit to the eventual x-ray structure.
Even x-ray structures, produced as they are at very low temperatures in the solid state may not reflect the 'real' situation in vivo, where the protein is in aqueous solution and may well be associated metal ions (esp. calcium) AND other proteins AND be in dynamic equilibrium with various conformations of itself.
So great if the number crunching can be massively speeded-up but in all other respects there is still a long way to go.
The problem is really basic science. It's possible now to homology model a modest single domain protein in a few hours even on a desktop workstation. Understanding how the current algorithms and model assumptions fail is MUCH harder. In the end a protein is a dynamic entity and that makes it all much harder.
A few years ago I was interested in a kinase enzyme. A xtal structure was available but the reality turned out that the protein was in dynamic equilibrium between ( at least ) two forms. One was equivalent to the xtal structure, the other was a form that could be activated to give the (unstable) working form - what the structure of that was ???.
The whole reason to model proteins is to make use of the information gained - a fast method of getting the wrong, non-physiological answer is useless on it's own. I'm very optimistic really but there are many problems to solve that don't depend on calculation speed.
I suspect they weren't simultaneously entangled with each other, which I believe was the case with the eight, so rather than giving the equivalent of 2^81 states (a BIG number), they give 2x81 states (not a BIG number), or something...
This is based on my limited understanding of quantum wiggly-jigglyness of course, so I stand to be corrected by someone who has read and understood the paper in question...
Anyone who is interested in doing this themselves I recommend starting out with wafer thin ham as your best bet. Don't get the real cheap stuff either, it'll just fall apart in your hands. After a lot of practice you can move on to other meats like turkey strips and even bacon. One day I hope to fold a duck out of duck.
This post has been deleted by its author
Their key points seem to be.
a) It's *programmable* (like a real computer), provided you can figure out how to describe the problem in "qubit" terms. Learning to think *in* those terms is likely to be easier for the yoof (if this takes off and does not become yet-another-technology-of-the-future, always was and always will be).
b) They have fabricated a chip (8 elements, each element of what seem to be 4 input qubits and 4 output qubits) full of quibits, which *mostly* work
c) You could have an *array* of such chips. It's *scalable*.
d) They are superconductive. Page 13 onward of the supplementary material describes it. It appears to operates at 20mK !. The process uses 4 Niobium layers and TiPt for resistors. Lines & spaces are 0.25 micrometres and the SQUID junctions are 0.6 micrometres (so a fair bit of room for reduction). This seems the closest anyone has got to actually *selling* a superconducting computer.
They claim it searches s solution space of 2^81 entries (which is pretty clever) partly by cycling the temperature.
Sadly a really hot cup of tea will result in a massive out of range error :(.
BTW A lot of AI seems to have been done on protein structure analysis (you've got X-ray images). This seems to have helped knock down initial analysis from *years* to hours. I don't quite see why heuristics cannot be used to constrain the search space and carve it up into multiple (but much *smaller*) search spaces to deliver a confirmation.
This post has been deleted by its author
It seems to me that all this Quantum computing malarky is just hi-tech divination.
Consider the theory that 2+2=5 because some(one) has do be doing the counting. Is this not the same as the quantum wavefront collapse as you make an observation? (I might have that bit slightly wrong).
Is it possible that all the solutions to all the problems are actually "out there" and and by doing some sort of "magic ritual" you just focus your conciousness to the correct one?
I'm off to get my yarrow stalks to try to predict the downvotes :)
Mine's the coat that only exists when you're looking for it...
Unfortunately. you have no idea what constitutes a physical process. I'm sorry that I cannot remember the correct terminology ATM but I think that you will find that everything that you think is external is actually an approximation. For example, when you look at a distant star the image you "see" only resides inside of your brain. It has been modified by the retina before being further processed and delivered to your conciousness. Consequently it has no objective reality.
See "Dark Star", "talking to the bomb" for education
http://www.youtube.com/watch?v=qjGRySVyTDk
YMMV
'In nature, proteins should normally fold themselves to a “ground state” – the lowest possible energy configuration for that particular combination of amino acids'
Yeah no. Proteins do not do NP-hard computation in finite time. That would be "a locally low energy state", not "the globally lowest energy state".
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
