Speeding Up Rosetta Mutant Energy Calculations #516
Unanswered
brycejoh16
asked this question in
Rosetta
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Hi Rocco or Rosetta Team,
Is there a way to speed up Rosetta energy calculations for many single and double mutants by avoiding redundant computations?
I’m currently calculating Rosetta energies for a large set of single and double mutants by running an independent FastRelax protocol for each variant. However, this is quite inefficient since most of the protein remains unchanged across variants. Is there a way to avoid recalculating energy terms for parts of the protein that are unaffected by the mutation? For example, can I cache the one-body and two-body energy terms for the wild-type structure, and then only update the energy calculations for residues within a certain distance (e.g., 8 Å) of the mutated positions? In other words, is there a way to reuse energy terms for parts (or residues) of the structure that haven’t changed, so I don’t have to re-relax and rescore the entire protein each time?
This is quite important since our team will be running potentially on the order of billions of FastRelax simulations. It is my impression that we can do something like below to only allow repack on relevant residues (for example within a 8 Å radius of the mutated positions), and could devise something similar for the minimizer, but energy terms will still be calculated for all residues- and nothing is cached.
As always, thanks for the help!
-Bryce
All reactions