[gmx-users] Energy minimization: problems with ramachanrad score

[dimitry.a.suplatov]

Hello.
I'm doing homology modeling. My model after MODELLER has Ramachandran
Z-score = 0.7 (that is OK). Then Im doing energy minimization with
steep, emtol=200 KJ/mol*nm. Fmax was converged to 1.82684e+02 in about
4000 steps. Backbone RMS compared to MODELLER model is 0.27nm. Z-score
is -3.5 and that is very poor.
Questions:
1. Could 4000 steps be too much for minimization? May be I should use
not more than 1000 and do not care about Fmax?
2. What minimization protocol do you prefer? What about using different
alorythms step by step?

Thank you.
Dimitry.