Dear Modeller users, I have constructed a homology model of my human protein starting with a template of the yeast variant. The first problem I faced is that both proteins are very different, with only 23% homology. To solve possible errors in the best modeller structure I also carried out a 6ns molecular dynamics asuming that minor changes would happen. But I'm not sure if the first structure is the best one. And the option to performing a molecular dynamics for each modeller structure is very time consuming, moreover the RMS backbone deviation for the first five or ten ones is very small (0.8 to 1.1 A). Here it's my question : What is the normal protocol in order to obtain a good enough final structure ? Can I assume that choosing the following structures and performing a molecular dynamics simulation I will get almost the same final stuctures ?

Thanks in advance for any suggestion.

Cristian Obiol Biological Systems Modelling and Drug Design Research Group University of Barcelona