I just started to use Rosetta and I am trying to model my protein with truncated sequence (First ~200 aa). I have pdb file of the full sequence and I would like to know the structure will look with first ~200 aa. Which protocol is better for me to use? Comparative modeling or Ab initio modeling? I am also curious about what that input native proteins is doing? Should I just provide my full sequence pdb file?
I assume you have experimental evidence or some biological reason to be interested in the 200 aa truncation. In most cases the first 200 aa will just fold as expected in the absence of the other domains. When they misbehave it tends to be aggregation not alternate folding; Rosetta can't help you there. If you have the whole structure already, then for most purposes all you need to do is cut off the extra residues and FastRelax what remains; it will probably be close. Maybe you would want to aggressively loop model any loops in the newly exposed surface, or run FloppyTail on the newly exposed terminus.
If something else is going on - like the first 200 aa are only stable in the presence of the rest of the protein, or the latter part of the protein is needed to help the first part fold - then you can certainly try ab initio to get guesses on alternate states. I don't think homology modeling will help you - if the perfect homolog has been solved already (you said you have the xtal of the full sequence) then HM will just return mostly that.
Thank you! The reason I am interested is due to a cryo-EM structure. I have a poor resolution (~20A) cryo-EM map of a filament and this 200 aa proteins could be the monomer protein. Do you think I can do an abinitio modeling with the guidance of the density? Do you happen to know how to do that?
I think the first thing to try is just modeling the 200mer into your density directly, yes. I don't know how to do that. You should re-post that as a separate question (I want to model this existing structure into this other density to see if it fits) and then we'll farm it out to the people who are likely to know.