Cindy Boer

80 | Chapter 2.1 Table 5: Association between protein coding variants identified by exome-sequencing and mJSW. Genetic variants tested Gene Non-synonymous STOP gained mJSW P-value* CHST11 3 - 0.05 COL12A1 23 1 1.00 DOT1L 13 - 1.00 FGFR3 12 - 0.64 GADD45B 1 - 0.91 GNL3 5 - 0.64 IGFBP3 2 - 0.28 KMT2A 8 - 0.40 NCOA3 7 - 0.24 PIK3R1 3 - 0.80 PTHLH 0 - - RUNX2 2 - 0.05 SLBP 1 - 0.16 SULF3 3 - 0.05 TGFA 1 - 0.04 UPK2 1 - 0.74 *Burden Test P-value: burden test was performed between the protein-coding variants in the candidate genes and mJSW. The fact that we were able to identify six loci with the current sample size (13K individuals in the discovery) indicates that cartilage thickness is a phenotype providing a better yield in number of discoveries than the efforts ran with traditional composite radiographic scores. As a comparison, the largest GWAS study up to now, arcOGEN with 7,4K cases and 11K controls as discovery, yielded one locus in the overall analysis, and seven additionally in a number of stratified analyses. Interestingly, in the current manu- script we report on rs10948155, which is in high LD (r 2 >0.8) with a locus from arcOGEN which was only marginally associated (p-value below genome-wide significance threshold) with OA in males only[2]. By using a cartilage specific endophenotype, ev- idence for this locus is elevated here to genome-wide significance in the total popula- tion, underscoring the increased power when more specific endophenotypes are used. Endophenotypes are quantifiable biological traits intermediate in the causal chain be- tween genes and disease manifestation (in this case osteoarthritis). JSW can be pre- cisely measured throughout the life of individuals[7] and also displays variation in nor- mal subjects. Therefore, mJSW may be more tractable for the genetic dissection of OA. Across the cohorts in this manuscript, mJSWhas been measured in different ways, using both hand measured JSW on radiographs as well as (semi) automatic software which could have added some noise to the overall meta-analysis. Future cross-calibration of JSW measurements might aid in a more precise measurement and additional power to pick up genetic loci.