Hi,
I'm fairly new to using Rosetta suite, so sorry if silly questions. I'm running ddg_monomer app for a protein of about 150 aa.
A) the pdb outputs I get are all the same, there is no backbone or sidechain movement neither for wt nor for mut structures (50 for each as recommended). I intended to carry out the protocol of row 16 in Kellogg, 2011. I attached the options file, if you'd be so kind to take a look at it. What am I missing here?
B) I'm working in a 2GB RAM PC. After a while, I got the process Killed (just that in shell prompt), and I guess it was due to memory overload, I checked memory consumption. The process starts over 45% memory usage and ramps to 65% in a half hour when the process stops without completion. Can anyone tell me what I'm doing wrong, or I just need a better PC?
| Attachment | Size |
|---|---|
 options.txt | 611 bytes |
I have already run the app in a 4Gb RAM PC and it worked, but still no conformational search is done, the 50 structures remain the same. I have tried changing different options without success.
Did you have ddg_predictions.out and mutant_traj*** files outputed? What does your mut.mutfile look like?
Hi deltag,
Thanks for your reply, sorry to read it so late. The answer is YES, I have those outputs (noted that the mut_traj file is empty). Thats the reason it took me some time to come back, I trusted the ddg_predictions output and moved on, but I am still troubled with the fact that there is so litle deviation among the output structures, my concern is that the conformational search is not properly done in the way I setted the program. I used the long version of the mutfile:
total 2
1
H 55 D
1
G 78 L
The output mutant structures are identical to the wt, except in the mutated residue. Neighboring residues are not disturbed in any way, wich is very unusual.
I found that when using -ddg:dump_pdbs false (which does not write pdb outputs) memory consumption is far lower, perhaps because the structure information is stored transiently (in RAM) when calculating. Hope this can help someone.
I found that when using -ddg:dump_pdbs false (which does not write pdb outputs) memory consumption is far lower, perhaps because the structure information is stored transiently (in RAM) when calculating. Hope this can help someone.