Modelling the Transcriptional Regulation of Androgen Receptor in Prostate Cancer

Synopsis

Transcription of genes and production of proteins are essential functions of a normal cell. If disturbed, misregulation of crucial genes leads to aberrant cell behaviour and in some cases, leads to the development of diseased states such as cancer. One major transcriptional regulation tool involves the binding of transcription factor onto enhancer sequences that will encourage or repress transcription depending on the role of the transcription factor. In prostate cells, misregulation of the androgen receptor, a key transcriptional regulator, leads to the development and maintenance of prostate cancer. Androgen receptor binds to numerous locations in the genome, but it is still unclear how and which other key transcription factors aid and repress AR-mediated transcription. Here we measured the transcriptional activity of 4139 putative AR binding sites (ARBS) in the genome with and without the presence of hormone using the STARR-seq assay. Only a small fraction of ARBS showed significant differential expression when treated with hormone. To understand the underlying essential factors behind hormone-dependent behaviour, we developed both machine learning and biophysical models to identify active enhancers in prostate cancer cells. We also identify potentially crucial transcription factors for androgen-dependent behaviour and discuss the benefits and downfalls of each modelling method.