force point mutation with fixbb

[Anonymous]

I have run relax on a protein of ~100 aa (Rosetta 3.4). All residues show score below 0.0 except one LEU:

label fa_atr fa_rep fa_sol fa_intra_rep pro_close fa_pair hbond_sr_bb hbond_lr_bb hbond_bb_sc hbond_sc dslf_ss_dst dslf_cs_ang dslf_ss_dih dslf_ca_dih rama omega fa_dun p_aa_pp ref total
LEU_27 -3.80035 0.405712 0.674211 0.0228272 0 0 0 0 0 0 0 0 0 0 0.972498 0.171944 1.80864 0.319658 -0.1 0.475143

The positive value comes mainly from fa_dun. I carried out a fixbb using resfile so that all other residues are kept (but repackable, using NATAA). and the LEU_27 is allowed to mutate ( using ALLAA in resfile). But fixbb returns the same sequence. When I review the individual scores, the same situation persists. Only LEU_27 got positive score, but worse :

LEU_27 -6.96497 0.713341 1.1435 0.0456544 0 0 0 0 0 0 0 0 0 0 1.945 0.343887 3.61729 0.639316 -0.2 1.28302

No matter how many decoys I return, I always got the same structure. Using -ex1:level 4 and so on does not change the situation. How can I set fixbb to fix the positive value? Or should I manually mutate the residue? Your suggestion is welcome!

[Anonymous]

Is the score of the whole pose getting better? I mean, rare Dunbrack rotamers do exist in protein structures. They’re penalized because they’re uncommon, but if they’re allowing a great fa_atr for this and residues nearby, such that the whole pose energy decreases, then Rosetta is doing what it is designed to do.

Per-residue energies are worth considering, but a single positive energy is not concerning, especially when there is no evidence of a clash (you have fa_rep < |fa_atr|).

[attesor]

Thanks a lot for your comment, smlewis!
The whole pose did get marginally better RE score (from -201 to -204).
But in this case, the LEU seems to be the only black sheep with positive score. I just suspect that the whole score may be further lowered by fixing this residue.
Is it possible to let Rosetta output all debug info (energy score) when mutating this residue to all other possibilities?

[Anonymous]

There isn’t a straightforward way to get the information you want that I can think of. You should try making 20 resfiles for fixbb, each with PIKAA for each of the 20 types at your position of interest. That will force Rosetta to print the energies of all residue types at that position as you seem to want.

I think the pmut_scan code does this too.

[Author]

Posts