PEDMED is a lean Perl script for hypothesis generation/validation. It looks for enriched protein domains in a microarray dataset to help posit or reinforce a prediction of structure, mechanism, relatedness etc. in observed probeset markers.
As an input, PEDMED takes microarray expression data a scientist has already statistically analyzed and designated into groups of foreground (statistically significantly enriched) probesets and background probesets. The input file also has the RefSeq protein identifiers corresponding to the probesets.
PEDMED searches the NCBI Conserved Domain Database for each predicted protein corresponding to a probeset in the input file. PEDMED then counts each type of protein domain by superfamily and conducts a chi squared analysis using the foreground as the observed frequencies and the background as the expected frequencies. FDR is corrected by implementation of the Benjamini and Hochberg algorithm.
PEDMED’s results show statistically significantly enriched protein domains in the foreground relative to the background and the genes that correspond to the probeset markers containing those enriched domains.
License the PED-MED Software
Academic and non-profit licenses are available – please use the form below. For information about commercial licensing opportunities, please contact Stephen J. Susalka, PhD, at firstname.lastname@example.org. For general questions regarding this software, please contact Eric Jimenez.
Eric Jimenez & Lance Miller, Ph.D.