ADAN: a database for prediction of protein–protein interaction of modular domains mediated by linear motifs

Abstract

Motivation: Most of the structures and functions of proteome globular domains are yet unknown. We can use high-resolution structures from different modular domains in combination with automatic protein design algorithms to predict genome-wide potential interactions of a protein. ADAN database and related web tools are online resources for the predictive analysis of ligand–domain complexes. ADAN database is a collection of different modular protein domains (SH2, SH3, PDZ, WW, etc.). It contains 3505 entries with extensive structural and functional information available, manually integrated, curated and annotated with cross-references to other databases, biochemical and thermodynamical data, simplified coordinate files, sequence files and alignments. Prediadan, a subset of ADAN database, offers position-specific scoring matrices for protein–protein interactions, calculated by FoldX, and predictions of optimum ligands and putative binding partners. Users can also scan a query sequence against selected matrices, or improve a ligand–domain interaction.