RIDDLE performs superiorly in matching disjoint gene subsets. (a-c) RIDDLE matches many disjoint subsets of various gene sets created by dividing KEGG pathway sets (a) and GO biological process sets (b) randomly into non-overlapping halves, and splitting GO biological process sets (c) by those annotated prior to and after 9 February 2007. Note the failure of hypergeometric for all three cases and the incremental improvement obtained by each component and the final combinatory RIDDLE method.
Wang et al. Genome Biology 2012 13:R125 doi:10.1186/gb-2012-13-12-r125