Image: Mostphotos
PhD project We use mathematical and physical models to understand how the structure of DNA affects biological processes.
The research investigates how proteins quickly find important DNA segments for gene regulation, despite the genome being densely packed. Hi-C data, network theory and deep learning are used to identify important DNA regions and their role in nuclear processes.
Several processes in the cell nucleus, such as gene regulation, only start when certain key proteins bind to short segments of our DNA. The interesting thing about this process is that these segments can be many millions of times shorter than the entire length of our genome, which is tangled and clumped together at many different levels. So the question is - how can proteins find these segments efficiently?
Models that try to recreate how long it takes proteins to find certain parts of the genome grossly overestimate the search time compared to experiments. In my research, I use 'Hi-C' data, experimental data whose very essence provides an estimate of how the parts of our genome are in contact with themselves.
Using network theory, stochastic models and deep learning, I try to find significant regions in our DNA that are of great importance in this process, and understand how these regions affect the interaction between the processes of the cell nucleus.
