Understanding the functional relevance of regulatory elements and their associated genes is a major goal in the postgenomic aera. Recently, a comprehensive catalogue of human regulatory elements based on phylogenetic foot printing analysis has been established. This set of contains > 50% new motifs (105), which show no similarity to a binding site bound by a known factor. Here, we introduce a new method and analyse large scale genome wide microarray data together with 21921 putative promoters and predict the gene ontology categories of the regulatory elements contained in the catalogue. We are able to map 526 functions to 72 regulatory elements of the catalogue. We show that the new method greatly enhances the functional prediction power in the analysis pipeline. We rigorously validate our new method based on an unbiased "literature wide"-whole-Pubmed analysis of known transcription factors as well as based on random background models. We estimate that only 3% of our new predictions are false positive.