Predicting Antibody 'Commonness'

This project focuses on predicting antibody "commonness" using high-throughput single-chain sequencing data, with potential significant impact on infectious disease antibody development and autoimmune disorder research. The aim is to develop a computational method to analyze transcriptomic samples of B-cell receptor repertoires from multiple individuals, creating a statistical estimate for the likelihood of observing specific antibody lineages across populations. The research methodology uniquely combines large-scale antibody sequence data analysis with machine learning techniques to generate a "commonness" score for repertoire-derived lead candidates.
This project falls within the EPSRC Biological Informatics research area, leveraging computational methods to analyze complex biological data.
It further aligns with EPSRC's Artificial intelligence technologies developing state of the art generative and predictive models. The project involves collaboration with Roche, bridging academic research with industrial applications in drug discovery and pandemic preparedness.
The expected outcome includes developing a definition of 'commonness', with the ultimate goal of releasing an open-source model and weights for a general commonness predictor in sequence and 3d spaces, accompanied by a research papers.