Virtual Immune Cells – Combing Mechanism, Genomics & AI to Predict T cell responses

PROJECT OVERVIEW

Virtual cells are a new approach to address how to link single cell gene expression data with mechanistic biology of cells built on a mechanistic understanding of processes inside cells.  This hybrid-AI based approach brings together genomics, mechanistic modelling and machine learning to generate predictive models of complex biological processes. Immune responses to tumours and adaptive immune cell driven autoimmune diseases, both involve CD4 and CD8 T cell recognition of antigenic peptides expressed by MHC molecules, this process is key to driving the immune response. We have previously generated the POEM (Predictors Of Epitopes by Mechanism) model to provide insight into the classical MHC class I presentation pathway and predict with high confidence CD8+ T cell responses in tissues.

In this project we will develop a model for MHC class II and the MHC class I cross-presentation pathway to predict the milieu of cytokines and effector function of T cells occurring in anti-tumour immunity, autoimmune disease and infection.  We will utilise a plethora of datasets developing using proteomics (mass spec), spatial biology (spatial transcriptomics) and gene expression analysis to develop, parametrise and validate the model from tumour vaccine datasets (LungVax) and inflammatory disease (Taurus).

The aim of this project is to develop a new approach “virtual immune cells” to predict antigen specific immune responses in human tissues combining single cell biology and mechanistic biochemistry driving antigen processing and presentation.

KEYWORDS

AI/ML,

computational modelling,

autoimmune disease,

oncology,

mathematics 

TRAINING OPPORTUNITIES

The applicant will work on the intersection between systems biology and computational modelling. Training will be provided as required in python programming, machine learning algorithms, mathematical modelling, systems biology (single cell gene expression analysis) and immunology as required.

KEY PUBLICATIONS

Brown LV, McConnell M, Rosler R, Peiser L, Schmidt BJ, Ratushny AV, Gaffney EA, Coles MC, Applying population mechanistic modelling to find determinants of chimeric antigen receptor T-cells dynamics in month-one lymphoma patients, Immunotherapy Advances, Volume 5 (1), doi.org/10.1093/immadv/ltaf001, 2025

Bolton C, Mahony CB, Clay E, Nisa PR, Lomholt S, Hackland A, Chin PS, Smith CG, Alexiou V, Nguyen HD, Thyagarajan M, Shiekh Z, Davis P, Chippington S, Compeyrot-Lacassagne S, Davda S, Foley C, Turtsevich I, Ingledow B, Kupiec K, Kelly J, Hanlon MM, DiCarlo E, Jones LJ, Smith SL, Eyre S, Neag G, Kemble S, Madhu R, Palshikar MG, Korsunsky I, Gao C, Tran M, Dendrou C, Buckley CD, Coles MCC, Raza K,MAPJAG Study Group, Gravallese E, Filer A, Wei K, Eslam, Al-Abadi W, Rosser EC, Wedderburn LR Croft AP, Severity-associated molecular niches define the inflamed synovium in juvenile idiopathic 1 arthritis, Science Translational Medicine, 2025

Cosgrove J, Novkovic M, Albrecht S, Pikor NB, Zhou Z , Onder L, Mörbe U, Cupovic J, Miller H, Alden K, Thuery A, O’Toole P, Pinter R, Jarrett S, Taylor E, Venetz D, Heller M, Uguccioni M, Legler DF, Lacey CJ, Coatesworth A, Polak WG, Cupedo T, Manoury B, Thelen M, Stein JV, Wolf M, Leake MC, Timmis J, Ludewig B, Coles MC, B-cell Zone Reticular Cell Microenvironments Shape CXCL13 Gradient Formation, Nature Communications, 2020, Jul 22;11(1):3677. doi: 10.1038/s41467-020-17135-2

Cosgrove J, Alden K, Stein JV, Coles MC, Timmis J. Simulating CXCR5 Dynamics in Complex Tissue Microenvironments, Front Immunol 2021 Sep 7;12:703088. doi: 10.3389/fimmu.2021.703088 

THEMES

Systems biology, Immunology, Artificial Intelligence, Mechanistic Modelling

CONTACT INFORMATION OF ALL SUPERVISORS

mark.coles@kennedy.ox.ac.uk

tim.elliott@ndm.ox.ac.uk

gaffney@maths.ox.ac.uk

miles.weatherseed@immonc.ox.ac.uk