EastBio: New Machine-Learning approaches for modelling protein-RNA interaction dynamics in bacteria.

To survive in hostile environments, cells must swiftly adapt their transcriptome to counteract stresses. RNA-binding proteins (RBPs), including ribonucleases, play crucial roles in adaptation by rapidly remodeling gene expression profiles post- transcriptionally, which can be achieved more efficiently than switching transcription factors on or off. Unsurprisingly, RBPs are thus attractive targets for antimicrobial therapies and are also frequently associated with a variety of human diseases, including cancer and neurological disorders.

Cross-linking and Immunoprecipitation (CLIP) and related techniques such as CRAC have revolutionised the field of RNA biology by enabling researchers to map protein- RNA interactions in living cells at nucleotide resolution and at high-throughput. With recent advances in UV cross-linking we can now also obtain time-resolved CLIP/CRAC data, which enables us to study of how proteins interact dynamically with RNAs and how this impacts gene expression profiles. However, as with any high-throughput approach, these methods introduce a degree of noise that must be carefully quantified. Moreover, in analysing time-resolved data, we have only begun scratching the surface in mining all the biologically relevant information.

To address these important challenges, we will develop robust ML/AI approaches aiming to identify patterns in the time-resolved data that can provide biological insights into how RBPs contribute to shaping gene expression profiles. Specifically, we will investigate RNA-binding profiles from the bacterial pathogen S. aureus to better understand its adaptation to the human host environment, and ultimately, to identify druggable targets that can disrupt infection and virulence.

A main challenge will be to integrate data from various sets of experiments to simultaneously analyse the roles of multiple RBPs and extracting the relevant features for training ML/AI models. However, preliminary ML models generated by our team have already established proof of principle.

We are looking for motivated candidates to join our diverse, international, and multidisciplinary team, working at the cutting-edge of computational modelling applied to biomedical questions.

The project will be an exquisite opportunity for candidates with a background in computational biology (or similar) to develop valuable machine learning and interdisciplinary skills. If of interest, the project can also be an opportunity to gain additional lab experience.

Interested candidates are encouraged to get in touch with Dr Andrea Weisse (https://homepages.inf.ed.ac.uk/aweisse/) or Dr Sander Granneman (https://sandergranneman.bio.ed.ac.uk/).