Structure-informed direct coupling analysis improves protein mutational landscape predictions

Data repository for StructureDCA Data repository for the publication:Matsvei Tsishyn, Hugo Talibart, Marianne Rooman, Fabrizio Pucci. Structure-informed direct coupling analysis improves protein mutational landscape predictions. StructureDCA source code: GitHub Each subfolder contains its own README with a precise description of its content. Content DMS collections: - proteingym: MSAs, 3D structures, StructureDCA and other model predictions on the ProteinGym DMS dataset collection.- megascale: MSAs, 3D structures, StructureDCA and other model predictions on the MegaScale DMS dataset collection.- humandomains: MSAs, 3D structures, StructureDCA and other model predictions on the HumanDomains DMS dataset collection. StructureDCA use cases: - B1-beta-lactamases: Data and StructureDCA predictions for the NDM1 and VIM2 B1 metallo-β-lactamases for epistatic analysis.- ParE-ParD: Data and StructureDCA predictions for the ParD–ParE protein–protein interaction using a concatenated MSA.- Spike-ACE2: Data and StructureDCA predictions using RSA derived from the bound 3D structure of the Spike–ACE2 complex.- KRAS-DARPinK55: Data and StructureDCA predictions using RSA derived from the bound 3D structure of the KRAS–DARPinK55 complex.