Overestimated prediction using polygenic prediction derived from summary statistics

Abstract

Abstract Background When polygenic risk score (PRS) is derived from summary statistics, independence between discovery and test sets cannot be monitored. We compared two types of PRS studies raw genetic data (denoted as rPRS) the statistics for IGAP (sPRS). Results Two variables with high heritability in UK Biobank, hypertension, height, are used to derive an exemplary scale effect PRS. sPRS without APOE International Genomics Alzheimer’s Project (IGAP), which records ΔAUC ΔR 2 0.051 ± 0.013 0.063 0.015 Disease Sequencing (ADSP) 0.060 0.086 Accelerating Medicine Partnership - (AMP-AD). On rPRS performances hypertension assuming a similar size 0.0036 0.0027 (ΔAUC) 0.0032 0.0028 (ΔR ). For 0.029 0.0037. Conclusion Considering height Biobank sample results AD databases inflated. Independence well-known basic requirement studies. However, lot follow such requirements because impossible direct comparisons when using statistics. Thus, sPRS, potential duplications should carefully considered within same ethnic group.