PgmNr 158: Known disease variants in a population-wide analysis of 135,638 Finns.Authors:
H.O.H Heyne 1,2,3,4; S.M.L Lemmelae 1; J.K. Karjalainen 1,2,3,4; A.S.H. Havulinna 1; A.P. Palotie 1,2,3,4,5; M.J.D. Daly 1,2,3,4
View Session Add to Schedule
1) Finnish Institute for Molecular Medicine, University of Helsinki, Helsinki, Finland; 2) Program for Medical and Population Genetics, Broad Institute of MIT and Harvard, Cambridge, MA, USA; 3) Stanley Center for Psychiatric Research, Broad Institute of MIT and Harvard, Cambridge, MA, USA; 4) Analytic and Translational Genetics Unit, Massachusetts General Hospital, Boston, MA, USA; 5) Psychiatric & Neurodevelopmental Genetics Unit, Department of Psychiatry, Massachusetts General Hospital, Boston, Massachusetts, USA
Rare genetic variants influencing common or rare diseases often have large effect sizes with potential direct treatment implications; studying their effects comprehensively requires however large cohort sizes. The population-wide cohort of FinnGen offers an unprecedented opportunity to systematically investigate phenotype associations of rare variants because the founding bottleneck of Finland concentrates them to higher population frequencies.
Here, we studied genome-wide associations of 1,801 disease endpoints using nation wide electronic health record data of 135,638 Finnish individuals. These individuals carry a total of 1582 coding variants that are annotated as (likely) pathogenic (224 variants) or “conflicting interpretations of pathogenicity” (1358 variants) in ClinVar with phenotype associations in FinnGen (p-value < 0.01). We systematically compare those variants’ direction of effect and associated phenotypes between FinnGen and ClinVar. We validate known/disputed variant effects and find previously unknown phenotype associations. Of particular interest, the Finnish population bottleneck concentrates generally diverse recessive mutational spectra onto one or a small number of higher elevated frequency variants. This ‘Finnish disease heritage’ allows us to use FinnGen to examine the phenotypic consequences and departure from Hardy-Weinberg equilibrium for 777 variants (107 (likely) pathogenic) reported to be causes of severe recessive disease - lending empirical data to longstanding questions regarding selection in these scenarios. In heterozygous form, the 777 variants have more significant associations with phenotypes than random (p-value 10-9). We identify cases where heterozygous variant carriers display different or opposite-effect phenotypic associations from homozygous variant carriers – including a variant in SCN5A that has previously been found in recessive/compound heterozygous state in multiple individuals with severe cardiac arrhythmia. By contrast, the variant has a protective effect for cardiac arrhythmia/atrial fibrillation phenotypes (ICD codes I47/I48/I49, p=6x10-6, beta -0.63) in heterozygous carriers, which could, despite much lower frequency, be replicated in the UK Biobank (p=0.04, beta -0.39). By systematically investigating known disease variants in population wide electronic health record data we validate and refine disease associations while providing insights into pleiotropic variant effects.