Mendelnoulli, a longitudinal model, links diagnoses by age to polygenic scores and rare single-gene (monogenic) variants. It covered 348 diseases, 36 polygenic scores and 34 monogenic conditions in three biobanks. Against age-constant polygenic effects, it improved one-year discrimination for all 28 diseases tested (median AUC 0.95 vs 0.92). Flagging the top 5% in people without prior diagnoses, it found 8.6 more of every 100 new cases. It linked 27 of 30 genes to their known disease, ranking it first for 17 (7 expected by chance). The abstract reports no limitations.
Why it is interesting: In this model, polygenic effects on disease changed with age, and BRCA1 and BRCA2 differed in timing. It treats common and rare inherited variation together across 348 diseases.
Cite this item
BibTeX
Free sign-up. A digest of the new papers twice a week, grouped by domain.