Transcriptional regulation is the principal mechanism in establishing cell-type specific gene activity by exploring an almost infinite space of different combinations of regulatory elements, transcription factors with high precision. Recent efforts have mapped thousands of candidate regulatory elements, of which a great portion is cell-type specific yet it is still unclear as to what fraction of these elements is functional, what genes these elements regulate, or how they are established in a cell-type specific manner. In this dissertation, I will discuss methods and approaches I developed to better understand the role of regulatory elements and transcription factors in gene expression regulation.

First, by comparing the transcriptome and chromatin landscape between mouse and human innate immune cells I showed specific gene expression programs are regulated by highly conserved regulatory elements that contain a set of constrained sequence motifs, which can successfully classify gene-induction in both species. Next, using chromatin interactions I accurately defined functional enhancers and their target genes. This fine mapping dramatically improved the prediction of transcriptional changes. Finally, we built a supervised learning approach to detect the short DNA sequences motifs that regulate the activation of regulatory elements following LPS stimulation. This approach detected several transcription factors to be critical in remodeling the epigenetic landscape both across time and individuals.

Overall this thesis addresses several important aspects of cis-regulatory elements in transcriptional regulation and started to derive principles and models of gene-expression regulation that address the fundamental question: “How do cis-regulatory elements drive cell-type-specific transcription?”