Gil McVean, PhD, Frontiers in Human Genetics Conference, October 11, 2024

Longitudinal Modeling of Human Disease Comorbidities and the Implications for Our Understanding of Risk and Pathology

ABSTRACT

The analysis of longitudinal data from electronic health records (EHRs) has the potential to improve clinical diagnoses and enable personalized medicine, motivating efforts to identify disease subtypes from patient comorbidity information. Recently, we developed an age-dependent topic-modeling (ATM) method that provides a low-rank representation of longitudinal records of hundreds of distinct diseases in large EHR datasets and applied it to about 300,000 individuals from UK Biobank and over 200,000 individuals from the All of Us program. We defined subtypes of the 52 heterogeneous diseases (i.e., those that occur in more than one topic) based on their comorbidity profiles, finding differential genome-wide and locus-specific genetic risk profiles for at least 18 of these. Such stratification improves understanding of patient heterogeneity, leading to better identification of genetic risk, characterization of pathological pathways, and the discovery of new targets for medicines.