Seminars & Colloquia Calendar

William Bialek–Princeton University and The CUNY Graduate Center

Thursday, August 20, 11:00AM

Title
“Apparent action-at-a-distance in gene regulation"

This work is motivated by recent experiments that visualize the molecular events involved in controlling the read out of information encoded in DNA. In bacteria, specific proteins (transcription factors) bind to DNA very close to the point where enzymes bind to start reading out the genetic information and synthesizing (transcribing) mRNA, and it is easy to see how these binding events can interact. But in more multicellular organisms, the binding sites (enhancers) for transcription factors are very far from the sites where mRNA transcription begins (promoters). It has long been assumed that this distance along the DNA is closed by bending and looping, bringing the promoter and enhancer into contact. It now is possible to measure these distances directly, showing that the sites do come closer when the promoter is active, but proximity is not contact: there are still 100+ nm distances between interacting sites. These transcriptionally active sites are bathed in a condensed droplet of proteins, one of many examples of phase separation now seen throughout living cells. I will explain some of this experimental background, and then argue that there is an interesting statistical physics problem hiding in these data.

We can understand the apparent “action-at-a-distance” in transcriptional regulation if fluctuations in the internal structure or composition of the condensed droplet are able to transmit information over long distances. This would happen naturally if the droplet were poised near a critical point, or possibly near a first order transition. We have found a mechanism for self-tuning of finite droplets to conditions with long correlation lengths, so that there can be a sort of “self-organized criticality” even near equilibrium. Finally, the physical picture of a critical droplet predicts that the thermodynamic interactions among promoter and enhancers sites should lead to forces that displace these sites relative to one another in response to binding and activation events. These displacements should be measurable, quantitatively.

Joint work with Thomas Gregor (Princeton) and Gašper Tka?ik (IST Austria). A first account of these ideas is arXiv:1912.08579.