New deep-learning based tool capturing combinatorial mutations in SARS-CoV-2
The emergence of SARS-CoV-2 variants such as Beta, Gamma, Delta and Omicron is a critical factor for drug development. Researchers around Sai Reddy, head of the Synthetic Immunology lab, now present a computational tool that enables the selection and focused development of lead candidates of antibodies that have the most potential to maintain activity against a rapidly mutating SARS-CoV-2. The deep learning-guided approach identifies antibodies with enhanced resistance to the evolving virus.
Amongst the variants of SARS-CoV-2, Omicron (and it’s variants including the current XEC variant) in particular has challenged drug development because this version of the virus is characterised by numerous mutations in the so-called receptor binding domain (RBD) of the SARS-CoV-2 spike protein, i.e. the binding site that is targeted by therapeutic antibody candidates in order to neutralise the virus. But virus evolution is an ongoing process and new variants keep arising. Hence, the clinical life span of antibody therapies is short in virus infections.
Researchers from the Reddy-lab now developed a deep-mutational learning tool combining deep sequencing and machine learning. The quantum leap of this method: the team fostered not only single mutations in the receptor binding domain but created a comprehensive library of multiple and combinatorial mutations in the RBD.
The deep-mutational learning tool also predicts the impact of the RBD mutations on the biomolecular antigen-antibody mechanism telling the researchers if binding or non-binding is more likely to happen. Guiding antibody engineering during drug development, this tool has the capacity to predict binding and escape for a panel of therapeutic antibody candidates targeting diverse binding sites on the virus.
Find original article published in Nature Biomedical Engineering:
Frei, L, Gao, B, Han, J. et al. (2025) external page Deep mutational learning for the selection of therapeutic antibodies resistant to the evolution of Omicron variants of SARS-CoV-2. Nature Biomedical Engineering, https://doi.org/10.1038/s41551-025-01353-4
Learn about research in the Lab for Systems and Synthetic Immunology led by Sai Reddy.