{"database":"biostudies-literature","file_versions":[],"scores":null,"additional":{"omics_type":["Unknown"],"volume":["5"],"submitter":["Schottlender G"],"pubmed_abstract":["Structural similarity metrics such as the Tanimoto coefficient (TC) miss many functionally related compounds-indeed, 60% of similarly bioactive ligand pairs in the ChEMBL database show TC < 0.30, revealing a major blind spot that constrains ligand-based discovery. Our motivation is to overcome this blind spot and enable the recovery of structurally different yet functionally equivalent chemotypes that structure-based similarity fails to detect. Here, we introduce the bioactivity similarity index (BSI), a machine learning model that estimates the probability that two molecules bind the same or related protein receptors. Trained under leave-one-protein-out (LOPO) across Pfam-defined protein groups on dissimilar pairs, BSI not only outperforms TC but also surpasses modern molecular embedding "],"journal":["Frontiers in bioinformatics"],"pagination":["1695353"],"full_dataset_link":["https://www.ebi.ac.uk/biostudies/studies/S-EPMC12698616"],"repository":["biostudies-literature"],"pubmed_title":["Beyond Tanimoto: a learned bioactivity similarity index enhances ligand discovery."],"pmcid":["PMC12698616"],"pubmed_authors":["Marti MA","Schottlender G","Fernandez Do Porto D","Prieto JM"],"additional_accession":[]},"is_claimable":false,"name":"Beyond Tanimoto: a learned bioactivity similarity index enhances ligand discovery.","description":"Structural similarity metrics such as the Tanimoto coefficient (TC) miss many functionally related compounds-indeed, 60% of similarly bioactive ligand pairs in the ChEMBL database show TC < 0.30, revealing a major blind spot that constrains ligand-based discovery. Our motivation is to overcome this blind spot and enable the recovery of structurally different yet functionally equivalent chemotypes that structure-based similarity fails to detect. Here, we introduce the bioactivity similarity index (BSI), a machine learning model that estimates the probability that two molecules bind the same or related protein receptors. Trained under leave-one-protein-out (LOPO) across Pfam-defined protein groups on dissimilar pairs, BSI not only outperforms TC but also surpasses modern molecular embedding ","dates":{"release":"2025-01-01T00:00:00Z","publication":"2025","modification":"2026-05-26T17:07:41.868Z","creation":"2026-05-25T03:07:47.554Z"},"accession":"S-EPMC12698616","cross_references":{"pubmed":["41393791"],"doi":["10.3389/fbinf.2025.1695353"]}}