Despite the fact that we identified several new CNVs that are potential causal variants on trait-associated haplotypes, collectively these CNVs could explain less than 5% of previously reported GWAS hits. Nonetheless, these observations emphasize the need to consider all classes of variation (SNPs and all structural variants, common and rare) when fine-mapping causal variants within association intervals. Sequence insertions relative to the reference sequence represent a particular challenge for both fine-mapping and association studies, because their presence on an associated haplotype might be easily overlooked.
