Can artificial intelligence predict and design immunological synapses?

• Grant number: 24/12890-5
• Support Opportunities: Regular Research Grants
• Start date: October 01, 2024
• End date: September 30, 2029
• Field of knowledge: Biological Sciences - Biochemistry - Chemistry of Macromolecules
• Principal Investigator: Helder Veras Ribeiro Filho
• Grantee: Helder Veras Ribeiro Filho
• Host Institution: Centro Nacional de Pesquisa em Energia e Materiais (CNPEM). Ministério da Ciência, Tecnologia e Inovação (Brasil). Campinas , SP, Brazil

Abstract

Artificial intelligence (AI) has emerged as a transformative influence across various domains, impacting our daily lives. In the life science field, AI has played a pivotal role in pushing the boundaries of fundamental scientific understanding. Noteworthy breakthroughs include AlphaFold's success in predicting protein folding. Beyond fundamental research, AI applications extend to practical health science domains, demonstrating their potential in predicting the risk of diseases.In this proposal, our objective is to assess the capabilities of AI-based methods in predicting and designing immunological synapses, which are mediated by the interaction between T cell receptors (TCRs) and antigenic peptides presented by major histocompatibility complexes (pMHC). This interaction triggers the activation of T cells, orchestrating immune responses to combat tumor and infected cells.Predicting whether a TCR can effectively bind to a particular antigen presented by the MHC has far-reaching implications. It spans from advancing fundamental scientific knowledge regarding the mechanism of TCR:pMHC interaction to achieving practical applications in diagnosing diseases by inferring the specificity of a TCR sequenced from patients against target antigens. Furthermore, the success in predicting TCR:pMHC recognition can produce broad impacts on immunotherapy, allowing the prioritization and selection of T cells to be used as immunotherapeutic to target, for instance, cancer cells. To achieve this goal, we will leverage the increasing amount of available TCR and peptide sequencing experimental data, along with 3D structure data, to train accurate and generalizable classification and generative AI models. We will integrate the latest AI architectures with biological and structural information about TCR interaction interface. Additionally, we will utilize antigen representations to develop a cross-reactivity oracle capable of detecting antigen similarities for TCR-based vaccine design. (AU)