Automatic semantic meta-analysis of published proteomic papers relaibly identifies the most frequently reported differentially expressed proteins.