Computational identification of disease models through cross-species phenotype comparison

Abstract

The use of standardised phenotyping screens to identify abnormal phenotypes in mouse knockouts, together with the use of ontologies to describe such phenotypic features, allows the implementation of an automated and unbiased pipeline to identify new models of disease by performing phenotype comparisons across species. Using data from the International Mouse Phenotyping Consortium (IMPC), approximately half of mouse mutants are able to mimic, at least partially, the human ortholog disease phenotypes as computed by the PhenoDigm algorithm. We found the number of phenotypic abnormalities in the mouse and the corresponding Mendelian disorder, the pleiotropy and severity of the disease, and the viability and zygosity status of the mouse knockout to be associated with the ability of mouse models to recapitulate the human disorder. An analysis of the IMPC impact on disease gene discovery through a publication-tracking system revealed that the resource has been implicated in at least 109 validated rare disease–gene associations over the last decade.

Translating multiscale research in rare disease

“…Another significant development for interpreting gene–disease associations is the standardised phenotyping of mouse models performed by the International Mouse Phenotyping Consortium (IMPC) coupled with Mammalian Phenotype Ontology (MP). The study by Cacheiro et al (2024) applied the PhenoDigm algorithm comparing recent IMPC phenotype data to human disease using Human Phenotype Ontology. They were able to find matches for approximately half of the mouse mutants to their corresponding human ortholog-associated disease.…”