Abstract
PrePPI is a structure-based pipeline that predicts protein-protein interactions (PPIs) between two structured domains and between structured domains and short linear motifs (SLiMs) on a proteome-wide scale. Since the 2023 Computational Resource Issue of JMB, the PrePPI website has been significantly expanded and redesigned. The resource now includes interactomes for human, yeast, and E. coli proteomes with 3D models for high-confidence domain-level complexes and PDB templates for most of the SLiM-mediated predicted interactions. A key new addition is derived from the clustering of the PrePPI interactomes based entirely on the structure-based likelihood of an interaction. Remarkably these clusters exhibit functional coherence and provide an unprecedented proteome-wide depiction of the subnetworks of PPIs that underlie biological phenomena. The new website - https://honigcomplab.c2b2.columbia.edu/PrePPI - provides convenient access to these clusters, to structural models for each pairwise complex, and to function annotations for individual proteins, enabling multiple modes of biological discovery.