2
$\begingroup$

I have recently begun using Folding@Home and I am curious how people are prevented from cheating the system. It seems to me that unless the final result is easily verifiable users could submit bogus folds in order to quickly gain credits. Is it in fact the case that the final results of calculating a protein fold are easily verifiable?

$\endgroup$
  • 3
    $\begingroup$ What I think might be happening is the same set of sequences is sent to multiple computers, and the incoming results are compared a number of times before they're declared valid. So, for example, the same sequence could be sent to 50 computers, and a certain threshold need to match (say, 20 of them) before they're marked as valid. Then, if any other computer submits a different result, it's marked as invalid and not counted. $\endgroup$ – MattDMo Jul 27 '14 at 0:33
  • 1
    $\begingroup$ That is what happens. These distributed computing platforms send out duplicate jobs for computation and the results compared after all are returned. It is also possible to get a sense of confidence in the fold prediction by comparing sequence homology to what is already known for solved protein structures. $\endgroup$ – user560 Jul 31 '14 at 1:15
3
$\begingroup$

Besides duplication as verification there are numerous other computational methods to ensure that the result is valid. Proprietary software is used for folding at home so one can safely assume that they check hashes on processors and files regularly. If integrity is broken (by a crash, memory corruption or God forbid injection) the software client side and server side trashes the result. One can imagine that faulty hardware or other problems can result in a ban from the servers.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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