Writer: Maria Pathrakorn Kuratana
Introduction
Artificial Intelligence (AI) has been rapidly rising in development and usage, with notable AI models from industry increasing from 60% in 2023 to 90%. Developments for the biological academia field include technology such as AlphaFold that uses an amino acid sequence to predict a protein’s 3D structure, helping researchers identify protein structures which would’ve needed crystallography instead. Growth for this field continues to develop and is happening close by. Within the Faculty of Life Sciences at UCL, a team of researchers has developed an AI tool, ImmunoMatch, that predicts which antibody components naturally pair together.
ImmunoMatch – What is it, and how does it work?
ImmunoMatch is an AI tool developed to help predict and understand the assembly of antibodies in the human body, allowing the speeding of therapeutic antibody design.
Antibodies are “a protein component of the immune system that circulates in the blood, recognises foreign substances like bacteria and viruses, and neutralises them”. They are incredibly important, as they are key players in immune defence and are used as therapeutic drugs. ImmunoMatch is able to correctly distinguish heavy-light pairs from random pairings and learns features associated with different chain types. It can then be applied to spatial sequencing data to reconstruct antibody pairing when direct pairing information is not available.
Strong advancements in immunological research can be boosted by ImmunoMatch, where underlying rules of antibody assembly in human B cells can be revealed. Improvement in the computational design of antibodies and validation of therapeutic antibodies as a result can also potentially accelerate drug development, allowing for pharmaceutical progression to take place and plausibly leading to next-generation drugs. This technology could help scientists decode immune responses, design better antibody-based therapies, and even use it for personalised medicine.
Faces behind the paper
Co-authors Donjun Guo, Joseph C. F. Ng and Franca Fraternali are all part of the Fraternali Lab here at UCL, where they worked alongside Deborah K. Dunn-Walters, a Professor of Immunology and Associate Dean Research and Innovation at the University of Surrey.
Fraternali Lab is a bioinformatics and computational biology research group led by Professor Franca Fraternali that identifies molecular determinants in the functioning or mis-functioning of protein structures and protein-protein interactions. The group uses bioinformatic methods to analyse protein structures and interactions but also computational simulations to determine protein stability.
Other software and web tools have also been developed by Fraternali Lab to help with the support of bioinformatics and protein-protein interaction analyses, such as UNIPPIN: database with protein interaction data and BRepertoire: a web server that helps with the analysis of antibody repertoire.
Where is this industry going to head?
The rapid rise of AI is becoming an integral part of day-to-day life but also evidently for scientific research, as seen through ImmunoMatch. Its potential to improve antibody-based therapies and advance patient care is exciting to see for years to come, and we are intrigued to see the development research and technological output of Fraternali Lab.
kinesismagazine.com 