Help regarding protein modelling and docking. Can anybody provide me a protocol/gist of the steps involved to generate a pdb file from just a protein fragment?

Question

[maverick]

I am in the learning process of protein modelling and docking, so I am kinda lost involving a scenario where I have a protein fragment (sequence) and I need to develop a pdb file to continue with the docking procedures. However I am lost on how to generate a pdb file from just a protein fragment. A protocol or gist of the steps required would be highly beneficial. I have tried to use the I-TASSER server and did get a pdb file from it but I am suspicious the file is not accurate. So far from what I have read, I understand that I need to do homology modelling. Is that the correct method? Also is there another method which might be as feasible as going through the homology modelling procedures?

Answer 1

[jgreener]

I think by fragment you mean sequence - fragment usually has a different meaning, referring to a short structural motif (e.g. the structure of 5 amino acids from a larger structure).

If the protein has available templates then I-TASSER, SWISS-MODEL, PHYRE2 and others will all give you decent models. They should also report confidence measures to give you an idea of how reliable the models are.

If they fail to find templates you need to do de novo protein structure prediction, which is a little trickier. For this you could use Rosetta, CONFOLD2, or one of the recent deep-learning based approaches such as DMPfold from our lab. Doing this and then getting accurate docking results is quite ambitious though.

[maverick]

Thank you so much! I blasted my sequence and did get templates so I proceeded with SWISS-MODEL and I also checked my models in the SAVES server. Thank you again.