MadHitter_for_paper

MadHitter is a software package designed to find optimal combinations of gene targets for personalized cancer treatments. The program works by modelling the instances into integer linear programs and relies on a mixed integer programming software package to find optimal treatments for patients. Currently, we use SCIP or Gurobito solve the integer linear programs. The family of problems that we consider are variants of the hitting set problem from computer science literature. In our context, an instance is comprised of a cohort of patients, which corresponds to a collection of hitting set instances on the same universe. The universe is a set of genes. The elements of the hitting set instance for any patient are cells. Each cell may or may not express each gene; if the gene is expressed, then the level of expression is a numerical value. Thus, we say that a "gene g hits a cell c" if g has a non-zero expression value in c.