- This topic has 4 replies, 2 voices, and was last updated 12 years, 9 months ago by Anonymous.
February 9, 2011 at 12:10 pm #778Anonymous
I am running the script for protein design downloaded from http://pyrosetta.org/scripts.html.
It makes design of the last 10 aa of the protein while the rest of the protein remains unchanged.
When I calculated 100 decoys, in 99 cases I obtained the same sequence of the designed fragment (different from the native one).
When I allowed for changes of sidechains rotamers in the rest of the protein I also obtained sets of the same decoys.
It is strange… Is it ok???
Is it possible to monitor how the sequence is changing during design?
February 9, 2011 at 5:25 pm #4996Anonymous
The packer is designed to converge on the best solution when possible. This probably means the script poses an easy design problem, and so it can find the correct solution frequently.
Does it also place identical rotamers of those residues?
You can always turn the packing intensity up to make the search space larger, this may give more diversity in the results. You can do this by modifying the resfile; check the resfile documentation in the manual. (I have no idea how the sample resfile even works – it’s a Rosetta2 resfile, not a Rosetta3 resfile.
It’s also possible to turn the packer temperature up so that it converges less.
February 9, 2011 at 6:41 pm #4997Anonymous
Yes, the rotamers are also identical in many decoys.
In case when I allowed for sidechains rotamers changes in the rest of the protein, the sequences of designed fragment are identical in 84 decoys (from 100 calculated),
and among them there are only three unique decoys.
I was wondering if it is not the problem with random seed, that seems to be the same for different runs, but I do not know if it used here…
We had similar problem with some version of Rosetta2, which we used for de novo modelling. When we run it on the computer cluster, from some reason the random seed was the same for different runs and we were obtaining many identical decoys. To avoid it we had to specify random seed manually for each run.
February 10, 2011 at 1:12 pm #5002Anonymous
I checked with your server and in fact I obtain the same sequence for different random seeds, so it is not the case of random seeds…sorry
How I can change the packer temperature in PyRosetta?
February 10, 2011 at 3:11 pm #5004Anonymous
“I checked with your server and in fact I obtain the same sequence for different random seeds,”
This means you are getting a converged answer…this is generally a good thing, not a bad thing.
In Pyrosetta, see if you can access the PackerTask and see if you can get the high_temp and low_temp operations to work to change the packer temperature. Unfortunately the disallow_quench option is broken, so you can’t turn the quench step off.
- You must be logged in to reply to this topic.