Hi Greg Montr?al, I am the author of KEGGprofile package. Thanks for using my package. I think defining the reference genes based on your own experiments is the most reasonable method for finding overrepresentation pathways. For example, the high abundance proteins were easier to be identified in a proteomics experiment because of the limitation of current analysis method. If all of the proteins for the species were used as reference to find the overrepresented pathways for significant changed proteins, the pathways with high abundance proteins would be found. So I think the proteins identified in this experiment is better reference than all of the proteins for the species. For the find_enriched_pathway function, we used genes but not chip ids as input. So the function didn't know which chip was used in this experiment. I think a parameter that was used to define reference genes could be added in next version of this package. -- Shilin Zhao Key Laboratory of Systems Biology Shanghai Institute for Biological Sciences Chinese Academy of Sciences 320 Yue-Yang Road Shanghai,China,200031