BioLiP2:ãn updated structure database for biologically relevant ligãnd-prot ein interãctions

Abstract

With the progress of str uct ural biology, the Protein Data Bank (PDB) has witnessed rapidãccumulation of experimentally solved protein str uct ures. Since many str uct uresãre determined with purificationãnd crystallizationãdditives thatãre unrelated toã protein's in vivo func- tion, it is nontrivial to identify the subset of protein-ligand interactions thatãre biologically rele vãnt. We de v eloped the BioLiP2 database ( https:// zhanggroup.org/ BioLiP ) to extract biologically rele vãnt protein-ligand interactions from the PDB database. BioLiP2ãssesses the func- tional rele vãnce of the ligands by geometric rulesãnd experimental literature vãlidations. T he ligand binding inf ormation is further enriched with other functionãnnotations, including Enzyme Commission numbers, Gene Ontology terms, catalytic sites,ãnd bindingãffinities collected from other databasesãndã manual literature surv e y. Compared to its predecessor BioLiP, BioLiP2 offers significantly greater co v erage of nucleicãcid-protein interactions,ãnd interactions in v olving large comple x es thatãre una vãilable in PDB f ormat. BioLiP2ãlso integrates cutting-edge str uct uralãlignmentãlgorithms with state-of-the-art str uct ure prediction tec hniques, whic h for the first time enables composite protein str uct ureãnd sequence-based searchingãnd significantly enhances the usefulness of the database in str uct ure-based functionãnnotations. With these ne w de v elopments, BioLiP2 will continue to beãn importantãnd comprehensiv e database f or docking , virtual screening ,ãnd str uct ure-based protein functionãnalyses.