Rebonto Haque

I am a PhD researcher in Statistics at the University of Oxford (Magdalen College), supervised by Prof. Charlotte M. Deane at the Oxford Protein Informatics Group (OPIG). My research sits at the intersection of interpretable deep learning and protein sequence–structure–function relationships, with a particular focus on therapeutic antibodies and antibody–antigen interactions.

My work uses sparse autoencoders (SAEs) for mechanistic interpretability, and structure-conditioned generative models (for example, ESM-IF and finetuned variants) to study and design antibody sequences that preserve functional properties such as binding and cross-reactivity. I combine representation learning, generative modelling, and high-performance computing workflows to move from mechanistic insight to actionable sequence designs.

Supervision and partnerships

• Supervisor: Prof. Charlotte M. Deane
• Industrial partner: Leyden Labs
• Research group: Oxford Protein Informatics Group (OPIG)

Education

• DPhil (PhD) in Statistics, University of Oxford (Magdalen College), 2025–2029 (ongoing)
• MBiochem Biochemistry (Molecular & Cellular), University of Oxford (St Peter’s College), First Class Honours (10/110 in cohort), 2021–2025

Outside research I have served as President and Secretary of the Oxford University Chemistry & Biochemistry Society, led a winning team at the OxBioHackathon 2023, and contributed as a student consultant for the Oxford Strategy Challenge.

If you’d like to discuss collaboration, please get in touch at rebonto.haque@magd.ox.ac.uk.