PT  - JOURNAL ARTICLE
AU  - Adam Socrates
AU  - Tom Bond
AU  - Ville Karhunen
AU  - Juha Auvinen
AU  - Cornelius A. Rietveld
AU  - Juha Veijola
AU  - Marjo-Riitta Jarvelin
AU  - Paul F. O’Reilly
TI  - Polygenic risk scores applied to a single cohort reveal pleiotropy among hundreds of human phenotypes
AID  - 10.1101/203257
DP  - 2017 Jan 01
TA  - bioRxiv
PG  - 203257
4099  - http://biorxiv.org/content/early/2017/10/14/203257.short
4100  - http://biorxiv.org/content/early/2017/10/14/203257.full
AB  - Background There is now convincing evidence that pleiotropy across the genome contributes to the correlation between human traits and comorbidity of diseases. The recent availability of genome-wide association study (GWAS) results have made the polygenic risk score (PRS) approach a powerful way to perform genetic prediction and identify genetic overlap among phenotypes.Methods and findings Here we use the PRS method to assess evidence for shared genetic aetiology across hundreds of traits within a single epidemiological study – the Northern Finland Birth Cohort 1966 (NFBC1966). We replicate numerous recent findings, such as a genetic association between Alzheimer’s disease and lipid levels, while the depth of phenotyping in the NFBC1966 highlights a range of novel significant genetic associations between traits.Conclusions This study illustrates the power in taking a hypothesis-free approach to the study of shared genetic aetiology between human traits and diseases. It also demonstrates the potential of the PRS method to provide important biological insights using only a single well-phenotyped epidemiological study of moderate sample size (~5k), with important advantages over evaluating genetic correlations from GWAS summary statistics only.