Using a define structure as template for homology modeling

[Le-Phung-Hien]

Hi team,

Great work!

Just wonder instead of having to download the big Uniclust30 and PDB70, its it possible to run the "Multiple initial models mode" using one or two user-defined PDB crystal structures of known homologous as template?

Thank you very much.

Best regards,

Hien

[feiglab]

This could probably be done by tweaking the scripts but it is not clear that using just one or two PDB structures as a template will offer benefits over using a larger set of templates (or no templates at all, which is the original idea of PREFMD).

Also, if you have templates and AF2 builds confident models, those are probably more accurate than anything you may get via PREFMD refinement.