Todorov, K. | Search bioRxiv

bioRxiv2025

Therapeutic Monoclonal Antibodies Repurposing in Oncology via IMGT/mAb-KG Embeddings

BackgroundCancer remains one of the leading causes of mortality world-wide, accounting for approximately 9.7 million deaths in 2022. Faced with this significant public health challenge, therapeutic monoclonal antibodies (mAbs) have emerged as promising alternatives that may minimize the side effects associated with conventional treatments such as radiotherapy and chemotherapy. To support mAb research and development, IMGT(R), the international ImMuno-GeneTics information system, has established two standardized data sources namely IMGT/mAb-DB, a comprehensive database for mAbs, and, more recently, IMGT/mAb-KG, a dedicated knowledge graph for mAbs. Despite these advances, the development of therapeutic mAbs remains both time-consuming and financially burdensome--costs can reach up to $2.8 billion. To address this challenge and accelerate cancer treatment, mAb repurposing represents a promising alternative. ResultsIn this study, we leveraged a subset of IMGT/mAb-KG, dedicated to the oncology domain, to develop a scientific hypothesis generation application for mAb repurposing. This application, based on knowledge graph embedding techniques, is designed to suggest potential mAb candidates for novel oncology applications. A user-friendly web interface provides access to the tool, incorporating visual support to facilitate the interpretation of generated hypotheses. This application is a decision support tool aiming to accelerate the discovery of new therapeutic applications for existing mAbs. ConclusionOur application demonstrates the potential of knowledge graph embedding techniques in the oncology domain by enabling the repurposing of existing mAbs for new therapeutic uses. Using this tool, we have identified two novel mAbs, loncastuximab tesirine and glofitamab, both currently undergoing clinical trials for the treatment of chronic lymphocytic leukemia. This decision-support tool thus facilitates the discovery of new therapeutic opportunities by effectively repositioning existing mAbs for oncological indications, potentially accelerating the development of cancer therapies and addressing critical public health needs.

bioinformatics↗

bioRxiv2023

IMGT/RobustpMHC: Robust Training for class-I MHCPeptide Binding Prediction

The accurate prediction of peptide-MHC class I binding probabilities is a critical endeavor in immunoinformatics, with broad implications for vaccine development and immunotherapies. While recent deep neural network based approaches have showcased promise in peptide-MHC prediction, they have two shortcomings: (i) they rely on hand-crafted pseudo-sequence extraction, (ii) they do not generalise well to different datasets, which limits the practicality of these approaches. In this paper, we present PerceiverpMHC that is able to learn accurate representations on full-sequences by leveraging efficient transformer based architectures. Additionally, we propose IMGT/RobustpMHC that harnesses the potential of unlabeled data in improving the robustness of peptide-MHC binding predictions through a self-supervised learning strategy. We extensively evaluate RobustpMHC on 8 different datasets and showcase the improvements over the state-of-the-art approaches. Finally, we compile CrystalIMGT, a crystallography verified dataset that presents a challenge to existing approaches due to significantly different peptide-MHC distributions.

bioinformatics↗