It's supposed that quantum computers could help to find the native state of all human DNA proteins.

Does anyone know how to formulate the protein folding problem into a decision problem solvable either by quantum annealers or universal quantum computers?

It would also be interesting to know, how to model a decision problem for classical computers, like Folding@Home.

  • 2
    $\begingroup$ Welcome to Bioinformatics.SE! I'm afraid in its current state this question is very broad and may not be a good fit for this site in its current state. $\endgroup$ Jun 8, 2017 at 3:32
  • $\begingroup$ @DanielStandage thanks! ooo! ok understand! hopefully someone helps, since there is a big curiosity about this :D $\endgroup$
    – ncomputers
    Jun 8, 2017 at 3:40

2 Answers 2


Protein folding problem can be viewed as a minimization problem. One can use quantum annealing to perform the minimization. Running this on quantum computers would improve the performance since they can perform the tunneling directly.

In fact, quantum annealing was used (blog post) for lattice protein folding on the the D-Wave quantum computer (128 qubits). But this was an extremely simplified model, with only 40 discrete states.


I would expect that without any further insights into protein folding, a quantum method would look fairly similar to existing methods, but might have a bit more capability of exploring multiple states at the same time (without multitasking).

Here's the approach used by Rosetta, which is used in both the Rosetta@home distributed computing project, and the foldit distributed gaming project:

Rosetta's strategy for finding low energy shapes looks like this:

  1. Start with a fully unfolded chain (like a metal chain with its ends pulled).
  2. Move a part of the chain to create a new shape.
  3. Calculate the energy of the new shape.
  4. Accept or reject the move depending on the change in energy.
  5. Repeat 2 through 4 until every part of the chain has been moved a lot of times.

We call this a trajectory. The end result of a trajectory is a predicted structure.

  • 1
    $\begingroup$ wow that's surprising. I would have expected the protein to start folding as soon as one end of it is able to, which is probably the first end to leave the ribosome. The last amino acid probably hasn't even been attached yet, let alone available for moving and calculating the energy :) $\endgroup$
    – J.J
    Jun 8, 2017 at 22:24
  • 2
    $\begingroup$ That does happen to some degree, but the Rosetta [computer] model doesn't do this. A whole bunch of odd things happen with protein folding. For example, there are chaperone proteins that encourage folding in a particular fashion to get through low-energy humps to reach a higher energy state. The foldit researchers have found a way to model chaperones using a technique called "humans". $\endgroup$
    – gringer
    Jun 9, 2017 at 1:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.