Investigating pro-inflammatory vs pro-remission mechanisms across immune disorders: integrating single-cell/spatial multiomics & ML/AI approaches for patient-level translational insights

PROJECT OVERVIEW

Immune-mediated diseases (IMDs), such as inflammatory bowel disease (IBD) and rheumatoid arthritis, are characterised by chronic inflammation and a loss of tissue integrity and function. These diseases affect ~10% of the global population and their incidence has been rising. The discovery that anti-inflammatory biologics can ameliorate disease marked a new era in IMD treatment, but as many as 40% of patients do not respond to available therapies.

We have recently generated a longitudinal, single-cell therapeutic atlas of IBD and RA, contrasting patient biopsies obtained before and after treatment in drug responders and non-responders. These analyses have revealed specific cellular and molecular changes associated with non-remission, but also epithelial, glial and stromal signatures associated with pro-resolution.

These findings have implications for patient stratification, biomarker identification and improved treatment strategies. We are now extending these analyses to other treatment modalities, tissues and diseases.

The successful applicant will make use of available and emerging single-cell and spatially-resolved multi-omic datasets (that capture genetics, the transcriptome, the cell-surface proteome, the T- and B-cell receptor repertoires and the epigenome) to investigate pro-inflammatory and pro-remission mechanisms across diseases. In addition to performing analyses with available pipelines and tools (e.g. Panpipes, COMPASS), the applicant will have the opportunity to develop and apply generative machine learning (ML) approaches to better encapsulate the relationships between genes, enabling the investigation of regulatory gene network perturbations or other subtle behaviour that is not easily identifiable with standard data analyses.

The student will also be involved in developing and contributing to a platform for the user-friendly visualization of high-resolution metabolic data, including the incorporation of AI-powered language learning tools. Depending on student interests, a wet-lab component can be incorporated into the project with respect to tissue profiling by spatial transcriptomics, hyperplexed imaging, or CRISPR/Cas9-based editing screens/experimental validation, for example.

KEYWORDS

Immunology, single-cell, spatial transcriptomics, machine learning, artificial intelligence

TRAINING OPPORTUNITIES

The Kennedy Institute of Rheumatology is a world-class research centre, located in the University of Oxford’s Old Road campus, housing basic and clinical scientists and bioinformatics working on immunology and inflammation. This project will combine state-of-the-art omics, bioinformatics, and ML/AI approaches and the student will receive regular training and mentoring with respect to immunology and computational biology.

The student will join a vibrant postgraduate community at the Kennedy and will benefit from attending seminars delivered by world-leading scientists in the department and across the University; from public engagement opportunities; and from transferable skills and other training sessions including in entrepreneurship and innovation. The student will present their work at group meetings and national and international conferences. 

KEY PUBLICATIONS

Thomas T, et al. (2024) A longitudinal single-cell therapeutic atlas of anti-tumour necrosis factor treatment in inflammatory bowel disease. Nature Immunology

https://www.nature.com/articles/s41590-024-01994-8

https://www.ndm.ox.ac.uk/news/study-maps-out-next-generation-of-drug-targets-in-autoimmune-diseases

Curion F, et al. (2024) Panpipes: Pipelines for multimodal single-cell and spatial transcriptomic data analysis. Genome Biology https://genomebiology.biomedcentral.com/articles/10.1186/s13059-024-03322-7; https://github.com/DendrouLab/panpipes; https://panpipes-pipelines.readthedocs.io/en/latest/

Cui H, et al. (2024) scGPT: Towards Building a Foundation Model for Single-Cell Multi-omics Using Generative AI. Nature Methods https://www.nature.com/articles/s41592-024-02201-0

Bolton C, et al. (2025) Synovial tissue atlas in juvenile idiopathic arthritis reveals pathogenic niches associated with disease severity. Science Translational Medicine https://www.science.org/doi/10.1126/scitranslmed.adt6050?url_ver=Z39.88-2003&rfr_id=ori:rid:crossref.org&rfr_dat=cr_pub%20%200pubmed

THEMES

Inflammation

Computational biology

Machine learning

Clinical pathology 

CONTACT INFORMATION OF ALL SUPERVISORS

calliope.dendrou@kennedy.ox.ac.uk

dan.woodcock@nds.ox.ac.uk

stephen.taylor@well.ox.ac.uk