Why are the gene counts within various categories different between DAVID and EASE?
DAVID reports the number of accessions that belong to a given category, EASE reports the number of genes that belong to a given category. This difference occurs with accession systems such as Genbank that can have many accessions for the same gene. DAVID leaves the choice of what is to be counted up to the user; whatever is input is what gets counted. EASE, on the other hand, cannot simply use the accessions supplied by the user, since this might lead to a gene with multiple accessions receiving more that one “vote” in the over-representation analysis. So EASE converts all accessions to LocusLink numbers before reporting counts. This of course begs the question of whether a unique LocusLink number is the best way to define a single gene. This definition is not without problems, however we have found that it is superior to other methods such as using the gene symbol.
