In total, we studied 17,361 binary and 1,419 quantitative phenotypes. For all binary phenotypes, we matched controls by sex when the percentage of female cases was significantly different (Fisher’s exact two-sided P < 0.05) from the percentage of available female controls. This included sex-specific traits in which, by design, all controls would be same sex as cases. As a result, 10,531 (60.7%) of the binary phenotypes required down sampling of controls to match the case female percentage (Supplementary Table 1). Finally, to allow for more compartmentalized ICD-10 chapter-based analyses, all 18,780 binary and quantitative trait phenotypes were mapped to a single ICD-10 chapter including manual mapping for the non-ICD-10 phenotypes. Chapter mappings are provided in Supplementary Table 1. It is acknowledged that chapter mapping may have the greatest utility for diagnostic, rather than procedural, ICD-10 codes. For procedural codes, genetic associations could be incorrectly interpreted if chapter mappings are relied on. For example, surgical procedures commonly performed for patients with cancer are categorized within the dermatology chapter. Genetic associations reported for these procedures would be categorized within the dermatology chapter, but the underlying disease process is instead most probably reflective of an oncological aetiology.
