Runs the desired analysis on the smallest meaningful subset of the dataset. Traditionally, the smallest subset is either an individual locus or read.