Inference of transcriptional regulation from STARR-seq data.

Area of Science:

• Molecular Biology
• Genomics
• Computational Biology

Background:

• Transcription, regulated by RNA polymerase II (Pol-II) and transcription factors (TFs), is a key cellular process.
• Enhancers are cis-regulatory elements crucial for controlling transcription.
• Massively parallel reporter assays (MPRAs), like STARR-seq, enable large-scale measurement of enhancer activity.

Purpose of the Study:

• To develop a computational method for inferring the number, location, and width of functional regulatory sites within enhancer sequences.
• To identify local regulatory sites that may be overlooked by global computational approaches.
• To apply the method to analyze androgen receptor (AR) bound sequences relevant to prostate cancer.

Main Methods:

• A novel computational method based on a mean-field thermodynamic model of Pol-II binding, incorporating TF interactions.
• Application to simulated STARR-seq data to assess data quality impact on inference.
• Analysis of experimental STARR-seq data from AR-bound sequences.

Main Results:

• The method successfully infers functional regulatory sites within enhancer sequences.
• It demonstrates the impact of data quality on the accuracy of site inference.
• The approach identified key regulatory sites in AR-bound sequences, overlapping with known coregulator binding sites.

Conclusions:

• The developed method provides a powerful tool for dissecting enhancer function at a local level.
• It enhances the understanding of gene regulation by identifying specific functional sites within enhancers.
• The findings have implications for understanding AR-mediated gene regulation in prostate cancer.