Speaker: Johannes Nübler
Subject: Active polymer models for the 3D organization of chromosomes
Location: TU Delft, A1.100
Author: Kristian Blom
There it is, my final seminar report just before my trip to Japan starts. On the 3rd of July I visited a talk given by Johannes Nübler, postdoc at the MIT Mirny lab. Johannes works on the analysis of Hi-C and Micro-C data and modeling chromosomal organization in yeast, mouse, and human cells. His particular focus is on bridging large scale polymer models of chromatin with more fine-grained models of nucleosomes. He is interested in the role of active processes in chromatin organization, e.g. transcription and chromatin remodeling. In this talk we focused specifically on modelling Hi-C data.
The talk started with an introduction of Hi-C (High-throughput sequencing). With this method one can analyze the spatial organization of chromatin in a cell, by quantifying the number of interactions between genomic loci that are nearby in 3-D space, but may be separated by many nucleotides in the linear genome. A shortcoming of Hi-C is that when two sites are not in close contact, no additional information is provided about how far these sites are separated from each other. Hi-C data is usually visualized by a heatmap where the number of interactions between all the different locations of a genome is shown. In order to make sense of this data, a model has been created that simulates active polymer folding and recreates the experimentally obtained heatmaps. By comparing the experimental heatmaps with the modelled heatmaps, one can understand why certain features arise under specific circumstances.
After the introduction we looked at interaction heatmaps of entire chromosomes, characterized by repeating blocks of high/low interaction intensity in the vertical and horizontal direction. The blocks with high interaction intensity are called topologically contacted domains. A typical feature in these interaction heatmaps are the dense squares along the diagonal, which indicates that neighboring areas on the chromosome are relatively often in contact with each other. To simulate those features passive polymer folding was not enough, and therefore a loop extrusion model was introduced (Figure 1). In this model DNA folding arises due to an extrusion complex containing two subunits that attaches to the DNA, forming a small loop in the process. Thereafter the two subunits slide along the DNA in opposite directions, making the loop bigger. This mechanism actively folds the DNA, allowing far located parts on the sequence to come in closed contact with each other. During sliding the extrusion complex searches for a specific motif, which causes a protein called CTCF to bind to the DNA. When a subunit encounters a CTCF that is pointed towards the subunit, it will stop sliding. If the CTCF motif is pointed in the sliding direction of the subunit, it won’t be recognized and the subunit keeps sliding. The result of this behavior is that the pair of CTCF motifs at the end of a loop are pointing towards one another. One of the exciting results of the extrusion model is that loops formed by extrusion will be unknotted, allowing for easy access to the genetic information. Another very nice and cool result is that a collection of chromosomes that are intertwined will naturally segregate when extrusion takes place.
Although you had to have a bit of background knowledge in Hi-C to understand everything, I really liked the theoretical aspect of the talk. The fact that the model is very simple but produces very interesting results makes it a very powerful model. For the future it should be experimentally validated whether DNA looping is an active process involving some sort of extrusion enzyme, or a passive process that arises due to DNA-DNA interactions. For now it seems that the former comes closest to reality, but we have to wait for evidence to know the truth. Now it is time to go to Japan, cheers!