Cindy Boer

260 | Chapter 6 nisms, pathways and pathology of osteoarthritis[20-23]. Another strategy to decrease phenotype heterogeneity, and thus increase power, is the use of stratified osteoarthritis phenotypes. For example, grouping individuals based on specific osteoarthritis symp- toms (pain vs. no pain, the amount of pain), based on pathological characteristics (os- teophytes and joint space narrowing) or joint function (range of motion, disability) or using osteoarthritis endophenotypes: cartilage thickness or the shape of the joint or bone. An overview of all osteoarthritis associated stratified and endophenotype SNVs is given in Figure 2 , and variants mentioned in this thesis are highlighted. From (osteoarthritis) GWAS to (osteoarthritis) Gene Finding novel osteoarthritis, or osteoarthritis stratified/endophenotype associated loci, is not the sole purpose of a GWAS. The ultimate goal of performing osteoarthritis GWAS is: to identify associated genetic variation to (i) elucidate genes and biological pathways involved in osteoarthritis pathology. Clearly indicating further steps after GWAS analysis; to identify genes and pathways associated with osteoarthritis, arguably more important that just identifying osteoarthrosis associated genetic variation. How- ever, due to the fundamental design of a GWAS, only an associated genomic locus can be identified directly, never directly a disease associated gene or pathway, without addi- tional evidence. This path from GWAS associated SNV to disease gene is, unfortunately, neither simple nor straightforward. It is a time and effort consuming endeavour, requir- ing extensive post-GWAS analysis, interdisciplinary knowledge and collaborations. In Figure 3 an overview of the possible steps that need to be taken to get from osteoar- thritis GWAS to novel disease insight, is presented. All, or several, of these steps were performed for each osteoarthritis GWAS in this thesis ( Chapters 2- 4 ). Elucidating GWAS: Causal variant(s) In order to find an answer to the main GWAS question: how does the GWAS associated variant effects the target gene and how does this contribute to disease risk? First three major difficulties in GWAS interpretation need to be addressed: (i) what is/are the caus- al variants(s), (ii) what is the target gene (causal gene) and (iii) in which cell type or cell state does the causal variant exert its effect on the causal gene? First, due to the design of GWAS studies, not a single SNV is associated, but a whole linkage block of SNVs in high LD (r 2 ≥0.8), which can span a region over 100kb containing hundreds of linked SNVs. Any one or combination of these may exert the disease causing risk[24]. Although, it is not easy to identify a causal variant (or always necessary), identification of the causal variant can lead to identification of the causal