SAAMBE-MEM: a sequence-based method for predicting binding free energy change upon mutation in membrane protein–protein complexes

• DOI: 10.1093/bioinformatics/btae544
• Publication Date: 2024-09-07T01:10:33Z
• Authors: Prawin Rimal; Shailesh Kumar Pa...; Wang Xu; Yunhui Peng; Emil Alexov
• Keywords: Feature (linguistics)

Abstract Motivation Mutations in protein–protein interactions can affect the corresponding complexes, impacting function and potentially leading to disease. Given abundance of membrane proteins, it is crucial assess impact mutations on binding affinity these proteins. Although several methods exist predict free energy change due most require structural information protein complex are primarily trained SKEMPI database, which composed mainly soluble Results A novel sequence-based method (SAAMBE-MEM) for predicting changes (ΔΔG) complexes has been developed. This utilized MPAD contains affinities wild-type mutant complexes. machine learning model was developed ΔΔG by leveraging features such as amino acid indices position-specific scoring matrices (PSSM). Through extensive dataset curation feature extraction, SAAMBE-MEM validated using XGBoost regression algorithm. The optimal set, including PSSM-related features, achieved a Pearson correlation coefficient 0.64, outperforming existing database. Furthermore, demonstrated that performs much better when utilizing evolution-based contrast physicochemical features. Availability implementation accessible via web server standalone code at http://compbio.clemson.edu/SAAMBE-MEM/. cleaned database available website.