ChEMBL

One thing new in the backend and interface for this release of ChEMBL is the ability to search for targets containing particular PFAM domains. So if you know a PFAM id, you can search in the search box (and then select "Targets" for that domain. For example, PF00001 is the Pfam ID for the rhodopsin-like GPCRs.

A couple of important things on this though - the current functionality does exactly what it says - it returns proteins that contain that domain - the compounds do not necessarily (and often in fact do not) bind at that domain. This multidomain, and multi protein target issue is a surprisingly big challenge, and is a big trap for the unwary. So caveat emptor.

We do plan in the next release or two, provide a prediction of the likely/known compound binding domain (however here, for proteins that contain multiple copies of the predicted/lknown binding domain it is complicated....).