GSUB (Gwas-by-SUBtraction) is a command line tool to perform GWAS-by-subtraction. It uses closed-form solutions to estimate the associations between each SNP and the two latent components for two traits (i.e., $\gamma_1$ and $\gamma_2$ in the diagram below). Its inputs are two GWAS summary statistics for two traits, and the outputs are two summary statistics for the two latent factors. The $\gamma_2$ is shown in red is because it is the parameter of interest: the SNP effect on the latent factor for trait 2 after regressing out the factor for trait 1.

Please see the wiki for the detailed manual of GSUB.

• September 19, 2023: Initial release.

If you use GSUB, please cite:
Wu et al. (2023). Pervasive biases in proxy GWAS based on parental history of Alzheimer's disease. bioRxiv 2023.10.13.562272.

For questions and comments, please open a Github issue (preferred) or contact Yuchang Wu at [email protected] or Stephen Dorn at [email protected].