InterPro and Pfam: Protein domains and families for biomedical research

InterPro is a core protein data resource that amalgamates 13 protein family databases (including Pfam) to provide the definitive description and classification of protein domains and families. Various scientific communities rely on the range of annotations provided by InterPro and its member databases to acquire novel insights into the vast amounts of new DNA sequence data, enabling scientists to interpret experiments and design new ones based on annotations spanning complete genomes down to single residues. The past five years have witnessed a major increase in the scale of sequences, concomitant with the emergence of AI/ML approaches for unlocking the biological signals encoded within the data. This project will exploit these innovative approaches to enhance protein annotations and refine classifications through the incorporation of additional data types, along with the application of ML methods and organisation of the protein family data. We will integrate KOfams and AMRFinder into InterPro, while continuing to scale the InterPro/Pfam/HMMER resources to tackle billions of proteins. Cumulatively, these developments will accelerate biomedical research by facilitating scientists to decipher the effect of human mutations, understand drug interactions, combat antimicrobial resistance, and tackle emerging pathogens.