Technical Abstract: Peroxisome proliferator activated receptor (PPAR) agonists are currently used therapeutically in humans despite the fact that many of their direct gene targets are unknown. Since PPARs can directly regulate gene expression through peroxisome proliferator response elements (PPREs), we propose the computational prediction of human genes with upstream PPREs on a genome-wide scale. Based on our analysis, PPREs do not appear to be isotype- or species-specific nor do the nucleotides flanking the direct repeat confer additional PPRE discrimination ability. However, a position weight matrix (PWM)-based search for PPREs within upstream conserved elements yielded sufficient selectivity for a genome-wide search. Additionally, a novel motif that appears with greater prevalence than PPREs among the upstream conserved elements of reported human PPAR target genes is revealed. Microarray and gene ontology analyses further validate our search technique and provide new functional clusters of genes that were not previously known to be directly regulated by PPARs (e.g. chromatin remodeling). Of 24033 human genes, 1085 were found to contain PPREs in their upstream conserved elements. This first genome-wide library of high-confidence predicted PPAR target genes should be a valuable resource in the quest to dissect the complexity of PPAR biology.