Stacked neural network for predicting polygenic risk score.

In recent years, the utility of polygenic risk scores (PRS) in forecasting disease susceptibility from genome-wide association studies (GWAS) results has been widely recognised. Yet, these models face limitations due to overfitting and the potential overestimation of effect sizes in correlated variants. To surmount these obstacles, we devised the Stacked Neural Network Polygenic Risk Score (SNPRS). This novel approach synthesises outputs from multiple neural network models, each calibrated using genetic variants chosen based on diverse p-value thresholds. By doing so, SNPRS captures a broader array of genetic variants, enabling a more nuanced interpretation of the combined effects of these variants. We assessed the efficacy of SNPRS using the UK Biobank data, focusing on the genetic risks associated with breast and prostate cancers, as well as quantitative traits like height and BMI. We also extended our analysis to the Korea Genome and Epidemiology Study (KoGES) dataset. Impressively, our results indicate that SNPRS surpasses traditional PRS models and an isolated deep neural network in terms of accuracy, highlighting its promise in refining the efficacy and relevance of PRS in genetic studies.

Evaluation of optimal methods and ancestries for calculating polygenic risk scores in East Asian population.

Polygenic risk scores (PRSs) have been studied for predicting human diseases, and various methods for PRS calculation have been developed. Most PRS studies to date have focused on European ancestry, and the performance of PRS has not been sufficiently assessed in East Asia. Herein, we evaluated the predictive performance of PRSs for East Asian populations under various conditions. Simulation studies using data from the Korean cohort, Health Examinees (HEXA), demonstrated that SBayesRC and PRS-CS outperformed other PRS methods (lassosum, LDpred-funct, and PRSice) in high fixed heritability (0.3 and 0.7). In addition, we generated PRSs using real-world data from HEXA for ten diseases: asthma, breast cancer, cataract, coronary artery disease, gastric cancer, glaucoma, hyperthyroidism, hypothyroidism, osteoporosis, and type 2 diabetes (T2D). We utilized the five previous PRS methods and genome-wide association study (GWAS) data from two biobank-scale datasets [European (UK Biobank) and East Asian (BioBank Japan) ancestry]. Additionally, we employed PRS-CSx, a PRS method that combines GWAS data from both ancestries, to generate a total of 110 PRS for ten diseases. Similar to the simulation results, SBayesRC showed better predictive performance for disease risk than the other methods. Furthermore, the East Asian GWAS data outperformed those from European ancestry for breast cancer, cataract, gastric cancer, and T2D, but neither of the two GWAS ancestries showed a significant advantage on PRS performance for the remaining six diseases. Based on simulation data and real data studies, it is expected that SBayesRC will offer superior performance for East Asian populations, and PRS generated using GWAS from non-East Asian may also yield good results.

Employment and the sexual and reproductive behavior of female adolescents.

CONTEXT: Women's employment opportunities may reduce the risk of early intercourse and pregnancy, but some evidence has linked adolescent employment and problem behaviors with early intercourse. METHODS: Hazard regression analyses of data from the National Longitudinal Survey of Youth were used to examine the relationship between employment and the risk of first intercourse before age 20 among women who were aged 14-16 in 1979. The relationship between employment and the risk of a first, nonmarital pregnancy among sexually experienced young women was also assessed. RESULTS: Current employment and cumulative months of past employment are associated with increased hazards of first intercourse (hazard ratios, 1.20 and 1.01, respectively); this association is particularly strong for white young women. Adolescents who work more than 120 hours a month are significantly more likely than nonworking adolescents to experience first intercourse (1.4). Although current employment has no effect on the likelihood of a first, nonmarital pregnancy among white adolescents, it is associated with an increased risk of pregnancy among blacks and with a reduced risk of pregnancy among Hispanics. CONCLUSIONS: Program planners and policymakers should be aware of the potential association between adolescent employment, particularly intense employment, and the likelihood of initiating intercourse and experiencing pregnancy, even if causality is still unclear.