To do this analysis we intersected the 978 landmarks with a database of 1,533 gene sets compiled in Gene Ontology using the hypergeometric statistic (gene to GO gene ontology, conditional test for over-representation). We used the R Bioconductor package GOstats (v2.36.0) and the ontology from GO.db (v3.2.2). The results show that while some categories are enriched (e.g ATP binding, nucleoside/nucleotide activity, transcription factor binding, kinase regulator activity) the percentage of the 978 genes that are in any such set is small. While we did observe a number of classes to be enriched in the landmark genes, these categories tend to be generic (e.g. enzyme binding, protein kinase binding, catalytic activity, ATP binding) and/or contain only a small fraction of the landmark genes (e.g. protein kinase binding, which contains 84 of 978 landmarks). Taken together, we did not find any particular functional category dominating the list of landmarks chosen.
