Genome-wide association studies (GWAS) offer a powerful tool for identifying genetic variants for complex diseases, especially when large sample sizes are amassed. For diseases with limited sample sizes or for which case-only cohorts are available, public controls, who are not assessed for the disease, can be used without bias to cost effectively improve statistical power and novel locus discovery, if the disease prevalence is low in the general population1–5. Case-only GWAS datasets may exist for several reasons but primary among them is that the initial study focused on phenotypes within a patient population (e.g., set point viral load among those living with HIV6,7 or methadone dosing among those with opioid use disorder (OUD)8,9) but these case-only datasets could now be useful for GWAS of the primary disease (HIV or OUD) if paired with public controls. Combining cases and controls in this way is feasible even with samples genotyped on different array-based technologies10–13. A significant limitation of combining disease study cases with public controls is that unbiased results are only achieved using the intersecting set of variants genotyped across all arrays
