Disease-based Phenotypical Gene Prioritization

Wolpers, Anja and Helms, Tobias and Bohk-Ewald, Christina and Uhrmacher, Adelinde M. (2025) Disease-based Phenotypical Gene Prioritization. In: 23rd International Conference on Computational Methods in Systems Biology (CMSB 2025), 10-12 September, Lyon, France. Poster.

Image CMSBGenPrioAnja.pdf - Published Version Restricted to Registered users only Download (1MB)

Abstract

Diagnosing hereditary diseases using whole-exome sequencing (WES) is challenging because it is hard to identify the pathogenic genetic variant(s) among many benign ones. An established method for efficiently evaluating genetic variants is to refer to their genes and apply phenotypical gene prioritization. In phenotypical gene prioritization, genes are ranked based on how closely the diseases they are associated with match the patient’s symptoms, using the Human Phenotype Ontology (HPO). However, diseases often cause a wide range of symptoms, while individual patients usually show only a subset of those symptoms. While the HPO also includes information on the frequency of symptoms for many diseases, most existing algorithms do not exploit this information. Our approach to calculating the phenotypical similarity between a patient and a disease incorporates this symptom frequency information; i.e., frequent symptoms have a greater impact on the similarity than less frequent symptoms. Based on 6,735 pre-diagnosed clinical cases, our frequency-based approach crucially improves the results in 246 cases (4%). We therefore conclude that applying symptom frequencies can be a helpful tool to increase the effectiveness of phenotypical gene prioritization.