Two sets of CpG-islands were generated by means of the GenomeCluster program on each genome assembly: 1) the original set with p-value ≤ 1E-5, and 2) an strict set by increasing the required statistical significance to p-value ≤ 1E-20 (i.e. filtering out the less significant islands).

A track hub for each genome assembly was made available through the web, thus allowing 1) to display genome-wide maps of CpG-islands in the Genome Browser of the UCSC, and 2) to obtain tables with genome coordinates (in BED format) by means of Table Browser.