I need some help about the paper”Ultrafast shape recognition for similarty search in molecular databases”,in which a new method was introduced. I want to know if there is function in Pyrosetta to implement this method, if not I also want sth informatin about this question.
Thanks for your attention, and thanks for your help in advanced.(I am one Chinese, so forgive my poor english)
From what I can tell of the method, there isn’t. The method is straightforward enough, though. You could easily implement it yourself in PyRosetta, though I would recommend not bothering.
One of the issues you’re running up against is that (Py)Rosetta is intended mainly for the modeling of biological macromolecules, primarily proteins. The small molecule support that it has is there in service of modeling complexes with proteins.
The general principle when working with small molecules is to do all the protein independent chemoinformatics in external programs, and then use Rosetta for the protein-lignad interaction bits. Getting novel small molecules into Rosetta is a bit of a hassle, and slightly pointless if you’re not actually going to use the rest of the Rosetta machinery.
(All that said, I can see where the linked protocol might be useful for comparing protein structures. For example, by calculating the shape metric over all the Calpha atoms of a protein, rather than all the atoms of a small molecule. I’m unaware of it being used for that purpose, though.)