Surflex-PSIM is a local method to compare and align the binding sites of both homologous and divergent proteins. PSIM compares proteins based only on the surfaces they present in their binding sites, making use of both shape and electrostatic characteristics. The similarity definition is direclty relevant to ligand binding activity. Previous work has shown PSIM to have the sensitivity required to separate human kinases with different ligand binding profiles and the depth to discover 3-D motifs in extremely divergent ATP binding proteins. Here, we report two augmentations to PSIM: 1) the addition of an automated site detection method; and 2) a much faster alignment optimization method. Together, these improvements allow PSIM to function as a completely automated protein annotation tool. We will present the results of a temporal validation experiment in which PSIM was able to annotate the functions of a set of proteins whose structures were known prior to their established functions. We will also present preliminary data showing that PSIM can also annotate and predict off-targets for some small molecule ligands utilizing only binding site similarity.

Annotating protein-ligand relationships with binding site surface similarity using Surflex-PSIM

245th National Meeting (2013)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.