Some recent and not so recent projects:
Chromatin structure dynamics
In living cells the 3D structure of gene loci is dynamic, but this is not revealed by experiments such as chromosome-conformation-capture or fluorescence microscopy, which are performed on fixed samples. Although live cell imaging methods have advanced significantly in recent years, there is still significant gap in our understanding. To address this issue, we adapted the HiP-HoP polymer simulation scheme to study the configurational dynamics of a gene locus. Our previous work showed that the Pax6 gene locus can adopt a highly diverse range of structures within a population of cells; this is consistent with several experimental studies which showed broad heterogeneity in structural features such as topological domains and promoter-enhancer loops. How such variability can be compatible with a robust transcriptional programme and emerging phenotype is a remarkable phenomenon. Our simulations suggest that chromatin structures at the gene locus level are also highly dynamic in a single cell, and predict that the Pax6 locus will display all of its possible configurations within a time-scale of less than 1 hour. In other word, it suggests that the same level of variability will be observed within a single cell during a cell cycle, as is observed across the population.
with Giada Forte, Davide Marenduzzo and Nick Gilbert
Simulating cancer genomes
Many cancers arise due to genome rearrangements such as chromosome translocations, where two parts of different chromosomes can become fused together after a breakage. Genome regulation often involve chromosome looping and founding in 3D, and rearrangements can change the local 3D environment of a gene leading to mis-regulation. In this work we used polymer simulations based on the HiP-HoP model to study the changes in 3D structure which occur after a rearrangement. This has provided new insight into the processes involved in gene mid-regulation, and has also help us identify "interaction hotspots", which could in the future be targets for therapies. In our initial work we focused on rearrangement in B-cells involving the immunoglobulin heavy locus and the cell cycle gene CCND1. B-cells are part of the immune system, and programmed DNA breaking an repair is used to generate the genetic diversity required for generating immunity. Unfortunately these processes can go wrong, giving rise to different malignancies.
with Lisa Russell and Daniel Rico
Chromatin mediated phase separation of simple model proteins
Many proteins, including some involved in gene regulation, have been observed to form phase separated condensates, or liquid-like droplets. Inspired by some confusing experimental observations on heterochromatin protein 1 (HP1), we developed a simple course grained model for proteins with multiple domains which have different functions. By exploring the parameter space of protein-protein and protein-chromatin interaction strengths, we found several regimes with different interesting behaviour. Relevant for HP1 is a regime where the proteins form 'droplets' on chromatin, but display quite different behaviour to what would be expected in a liquid phase-separating system. Namely, the concentration of protein with the droplet varies with the overall concentration, while droplet size is buffered. This, and several other results have strong implications for our understanding of the role these processes play in vivo.
with Marco Ancona
The Highly Predictive Heteromorphic Polymer (HiP-HoP) Model
We developed a polymer-based simulation model for predicting the folding and looping of chromatin fibres at high-resolution around specific genes. The model combines several mechanisms which drive chromosome organisation (including the bridging-induced attraction and loop extrusion) with a heteromorphic polymer. Put simply, this is a polymer which has properties which vary along its length - in this case some regions have a thicker more compact structure, which others are thinner and more flexible. We have applied the model to study the Pax6 gene locus in mouse. A simpler version of the model was used to study the alpha and beta globin genes.
with Nick Gilbert and Davide Marenduzzo.
Chromatin domains in yeast
In recent experiments using MicroC (a nucleosome resolution HiC-like chromosome conformation capture method) revealed that the yeast genome is organised into domains of enriched self-interactions. These are, however, much smaller than the topologically associated domains (TADs) found in mammals and other organisms (5-10kbp domains in yeast compared to 100kbp-1Mbp in mammals). To study the mechanism of formation of these domains we developed a simple polymer model for chromatin where nucleosomes were represented by 10nm sphere, and linker DNA was represented by chains of 2.5nm spheres. Using data on nucleosome positions (from MNase-seq experiments) to set the lengths of DNA linkers allows specific chromosome regions to be simulated. Despite being highly simplified, the model gives very good predictions of MicroC interaction maps - since nucleosome positions are the only input, this implies that the small scale domains arise simply due to the underlying chromatin structure, and do not require the more complex mechanisms found in mammals.
with Oliver Wiese
Simulating Biofilms in marine environments
Surface-attached bacterial biofilms cause disease and industrial bio-fouling, as well as being widespread in the natural environment. We have developed several computational and analytical models for the establishment and growth of biofilms in different environments. In our model, individual, planktonic bacteria adhere to a surface, proliferate and undergo a collective transition to a biofilm phenotype. This predicts a stochastic transition between a loosely attached, finite, layer of bacteria near the surface, and a growing biofilm. To avoid industrial bio-fouling, antifouling surfaces are often used. By extending the model to include multiple species with different levels of biocide-resistance, our simulations can shed light on the factors affecting the efficacy and longevity of such surfaces.
with Rosalind Allen and Patrick Sinclair
P. Sinclair et al, Under Review (2022)
Many important cellular processes, such as gene regulation, require proteins to bind to short specific target sequences on large DNA molecules. The target needs to be found quickly and accurately. To do this it is thought that proteins move through the bacterial cell, or eukaryotic nucleus by making alternating rounds of free 3D diffusion, and 1D scanning along the DNA molecule. Whilst there is a long history of theoretical treatment of this process, there has been little work on simulations which take into account the full dynamics of both the proteins and DNA. Using coarse grained Brownian dynamics we have studied aspects such as the effect of DNA configuration, the presence of non-target "traps" in the DNA sequence, and the effect of crowding proteins in the cellular media, including those which are diffusing freely, as well as those bound to the DNA.
with Davide Marenduzzo and Mike Cates.
The bridging induced attraction as a driver of chromosome organisation
There are many DNA and chromatin binding proteins which from complexes which can mediate bridging interactions between different chromatin regions. Using coarse grained Brownian dynamics we uncovered a general tendency for such bridge-forming complexes to form clusters, even in the absence of attractive interactions between the complexes. This "bridging-induced attraction" not only provides a mechanism for the formation of the protein foci which are observed in vivo via microscopy, but it can also act to drive large scale chromatin organisation.
with Davide Marenduzzo and Peter Cook.
Ribosome Traffic in mRNA Translation
We use the versatile TASEP as a model for ribosome traffic flow in the biological process of protein production. This is a fundamental model in non-equilibrium statistical physics. We study features such as the effect of bottlenecks, and the interplay between supply and demand of resources, and how this might impact (and allow control of) protein levels and ultimately the characteristic of a cell.
with M Carmen Romano, Marco Thiel, Celso Grebogi and Ian Stansfield.