POLYGENIC RISK SCORES, PGT-P, AND IVF: SIBLING AND NON-EUROPEAN ANCESTRY VALIDATIONS

Polygenic Risk Scores (PRS) are under intense investigation by research groups all around the world. Large genomic datasets and machine learning methods have produced predictors for common disease conditions such as coronary artery disease, diabetes, schizophrenia, breast cancer. The objective of this study is to further validate PRS related health longevity indices, using sibling (family) cohorts non-European (African, South Asian, East Asian) ancestry groups. We analyzed effectiveness polygenic predict status among siblings. In previous work, were trained exclusively on European individuals, then validated a cohort consisting siblings similar genetic ancestry. analysis, in populations from ancestries. construct linear models predisposition variety conditions. condition phenotype data describes binary case-control which defined either self-report or clinical diagnosis. constructed compressed sensing. evaluate performance these (not used training) individuals present novel results several areas: 1. Validation family/sibling data, 2. 3. Construction evaluation an overall score index scores. Predictor primarily evaluated relative risk reduction (RRR): how much particular reduced rank-order selection PRS, compared random selection. RRR value varies specific group applied. give each major highlight we can say with high confidence that achieved through This translates into important practice differences, depending upon individual condition. For example, Asian comparable Europeans cancer, heart attack, prostate cancer; strong hypertension, others. has potential significantly reduce burden. reviewed analyses provide evidence significant