A simulation to analyze feature selection methods utilizing gene ontology for gene expression classification

Display Omitted HighlightsWe determine the conditions where Gene Ontology (GO) benefits Feature Selection (FS).We develop a simulation that generates gene expression datasets from GO.We define BCAL to quantify the notion of connectedness of a biological condition.If BCAL?0.696 then GO improves FS; if BCAL≤0.389 then GO does not improve FS.If 0.389