Cindy Boer

162 | Chapter 4.1 Results Genetic Architecture ‐ Identification of osteoarthritis risk variants We performed genome-wide meta-analyses for osteoarthritis across 13 international cohorts stemming from 9 populations ( Supplementary Table 1 ), in up to 826,690 in- dividuals (177,517 osteoarthritis patients). Two of the cohorts are of East Asian, and 11 of European descent. We defined 11 phenotypes encompassing all major sites for osteo- arthritis at both weight-bearing and non weight-bearing joints ( Supplementary Table 1 and Supplementary Table 2 ). We identified 223 genome-wide significantly associ- ated SNVs across osteoarthritis phenotypes using a threshold of p-value<1.3x10 -08 , to account for the effective number of independent tests ( Figure 1 , Table 1 ). Eighty-four of these variants have not been associated with osteoarthritis before. Using conditional analyses, we identified 100 unique and independent signals across osteoarthritis phe- notypes, 52 of which have not been associated with an osteoarthritis phenotype before ( Table 2 and Supplementary Table 3 ). For each of the 100 signals we define the lead SNV as the risk variant with the strongest statistical evidence for association. There was no evidence for multiple independently-associated SNVs at any signal. All lead SNVs have a minimumminor allele frequency (MAF) ≥0.01% in each cohort, and demonstrate the same direction of effect in two or more contributing studies. Six lead SNVs are cod- ing (all missense), 59 reside within a gene transcript and 35 are intergenic. We report here the first signals for spine (n=1) and thumb (n=2) osteoarthritis, and increase the number of SNVs for hand (five new, three known) and finger (three new, two known) osteoarthritis, as this is the first large-scale GWAS effort for these phenotypes. Of the 100 independent genome-wide significant SNVs, 90 are common (MAF ≥5%) and four are low-frequency (MAF <5% and ≥0.5%). We detected six rare variant associations (MAF 0.03% to 0.11%) with large effect sizes (odds ratio range 3.03 to 9.52) ( Table 2 and Supplementary Table 2 ); one previously reported and five new findings. The new rare variant associations all arise in a large extended family in Iceland, with some branches heavily burdened with osteoarthritis ( Supplementary Information ). The previously-reported hip osteoarthritis rs143083812 locus is a mis-sense variant located in the smoothened, frizzled family receptor (SMO) gene, for which we found association with total hip replacement (p-value=1.11x10 -11 , OR=3.30, 95% CI=2.34-4.66)[7]. All of the new rare variant associations reside in non-coding sequence and are driven by the Icelandic cohort, in which imputation accuracy is excellent, with support from either the UK Biobank or Estonian cohorts.