Ho, Vikki

Abstract Polygenic risk scores (PRSs), which quantify inherited susceptibility to complex traits and diseases, have emerged as valuable tools for risk stratification and precision medicine. Despite their promise, PRS developed on European cohorts often demonstrate substantially reduced predictive accuracy in non-European populations, due to differences in genetic architecture. The disproportionate representation of European ancestry cohorts in genome-wide association studies (GWAS) leads to inequitable deployment of PRS technologies across diverse populations. Here, we introduce PRANA (Polygenic Risk Adaptation via Neural-network Architecture), a deep learning framework that adapts an existing PRS developed on one population to other ancestries. Unlike methods that require large-scale GWAS in the target population, PRANA leverages pre-trained PRS models derived from European cohorts and adapts them using modestly sized cohorts from the target population. We evaluated PRANA on seven complex traits in South Asian, East Asian and Ashkenazi Jewish populations, as well as in selected smaller East Asian subpopulations where the scarcity of training data poses a particular challenge. PRANA mostly improved predictive performance of the baseline PRS models by 5%-20% in terms of effect size (β) and Nagelkerke’s R², and, in most cases, outperformed existing cross-ancestry multi-PRS approaches. These results highlight PRANA as a scalable and practical strategy to reduce disparities in genomic risk prediction and advance the equitable application of PRS in diverse populations.

Abstract Objective: The aim of the study was to quantify baseline estradiol (E2) and estrone (E1) concentrations according to selected patient characteristics in a substudy nested within the MAP.3 chemoprevention trial. Methods: E2 and E1 levels were measured in 4,068 postmenopausal women using liquid chromatography-tandem mass spectrometry. Distributions were described by age, years since menopause, race, body mass index (BMI), smoking status, and use and duration of hormone therapy using the Kruskal-Wallis test. Multivariable linear regression was also used to identify characteristics associated with estrogen levels. Results: After truncation at the 97.5th percentile, the mean (SD)/median (IQR) values for E2 and E1 were 5.41 (4.67)/4.0 (2.4-6.7) pg/mL and 24.7 (14.1)/21 (15-31) pg/mL, respectively. E2 and E1 were strongly correlated (Pearson correlation [ r ] = 0.8, P < 0.01). The largest variation in E2 and E1 levels was by BMI; mean E2 and E1 levels were 3.5 and 19.1 pg/mL, respectively for women with BMI less than 25 and 7.5 and 30.6 pg/mL, respectively, for women with BMI greater than 30. E2 and E1 varied by age, BMI, smoking status, and prior hormone therapy in multivariable models ( P < 0.01). Conclusions: There was large interindividual variability observed for E2 and E1 that varied significantly by participant characteristics, but with small absolute differences except in the case of BMI. Although the majority of participant characteristics were independently associated with E1 and E2, together, these factors only explained about 20% of the variation in E1 and E2 levels.