Answers are grounded in this paper and cite their sources.
Leveraging the curated, structured knowledge of phenotype-genotype associations available in model organism databases has proven valuable for predicting gene-disease associations in humans. While efforts to improve algorithms for semantic similarity using phenotype and disease ontologies for mouse and human are ongoing, significant gaps remain. Specifically, there are no comprehensive, validated mappings of equivalent terms from the primary phenotype ontology for the laboratory mouse (i.e., the Mammalian Phenotype Ontology [MPO]) and the human clinical phenotype terminologies (e.g., International Classification of Diseases [ICD] codes) used extensively in electronic health records (EHRs) and for genome and phenotype-wide association studies in human genetics research. Using two case studies, we demonstrate the potential to advance disease gene discovery by bridging EHR-based and laboratory mouse phenotype terminologies and outline the challenges that need to be addressed to realize this potential. We then propose using new methods, such as combining ontology-based and AI-driven similarity matching, to address limitations in previous approaches for mapping equivalent terms in mouse phenotype terminologies and human clinical terminologies; applying these methods, we provide results from preliminary mapping focused on cardiovascular phenotypes to demonstrate the feasibility of cross-ontology alignment and one-to-many mapping. We conclude that such methods, combined with expert manual review for validation and refinement, are needed to bridge the gaps currently limiting the translation of gene-phenotype knowledge from model organisms to the discovery and validation of phenotype-disease-gene associations in humans.
No takes yet. Share an insight, caveat, or question.
Feng et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: