Genome wide prediction of HNF4 alpha functional binding sites by the use of local and global sequence context

We report an application of machine learning algorithms that enables prediction of the functional context of transcription factor binding sites in the human genome. We demonstrate that our method allowed de novo identification of hepatic nuclear factor (HNF)4 alpha binding sites and significantly improved an overall recognition of faithful HNF4 alpha targets. When applied to published findings, an unprecedented high number of false positives were identified. The technique can be applied to any transcription factor.