Transcriptional Regulation: When 1+1≠2

One of the fascinating questions in biology is to understand how an identical genome can give rise to distinct tissues with different functions, for example, brain and muscle. A key role in selectively decoding the genome is played by transcription factors (TFs), which bind to specific DNA sequences to help specify if and how much of a gene is expressed in a particular tissue. In a simple scenario, binding of TFs near a gene would result in activation of gene expression whereas in the absence of binding the gene would not be expressed. One of the objectives of computational biology is to use the genomic sequence to predict where TFs bind and to both qualitatively and quantitatively predict which genes it regulates. In this chapter, we will discuss how the information encoded in the genome in the form of DNA can serve as a discreet code where combinations of As, Ts, Cs, and Gs specify which TFs can bind. Further, structural features of DNA can be read by proteins to influence their structure and fine-tune their activity towards target genes. In practice, predicting genome-wide binding patterns of TFs based on sequence is problematic and even when we know where TFs bind, all bets appear to be off regarding the effect of TF binding on the regulation of genes. At the moment it sometimes seems as if 1 + 1 ≠ 2 when studying gene regulation. However, this mostly reflects our lack of understanding of the signaling inputs that specify if a gene is activated and at which level it is expressed. For example, in this chapter we will discuss how taking the three-dimensional organization of the genome and the chromatin context in which these binding sites are embedded into account can improve the link between binding of TFs and the regulation of genes. Eventually, by adding more and more pieces of the puzzle, we hope to identify what is missing in our current equations to model gene expression.