One of the characteristic features of the Ensembl project has been to go beyond presenting raw data aligned to the genome sequence by also presenting high quality consensus biological predictions, generated from automatic analysis pipelines developed to use the raw data as evidence. Examples are the Ensembl gene build pipeline generating protein coding genesets and the Ensembl comparative analysis pipelines generating genetrees and orthology and paralog relationships. The Ensembl regulatory build is the latest such pipeline and provides automatic, evidence based annotation of potential regulatory regions within the human genome. The primary inputs are maps of open chromatin created by DNase I hypersensitivity mapping and covalent modifications of histone protein tails assayed by chromatin immunoprecipitation (ChIP). The first build was released in coordination with the ENCODE Pilot Project publication (9). Since the first release reported last year (14), we have updated the regulatory build three times, each time adding more data (35,36) and a more sophisticated analysis of the chromatin conformation and modification data. The build now consists of approximately 175 000 genomic regions defined from data collected from several
