our team studies the immune system as a dynamic, multi-layered control architecture and views tumour immunity as a departure from homeostatic set-points that can be measured, modelled and therapeutically restored. Alongside classical receptor-based checkpoints, we place particular emphasis on non-receptor mechanisms of immune regulation, including transcriptional and post-transcriptional regulation. Across disease settings we combine deep clinical sampling, multi-omic profiling, structural biology, and ex vivo functional platforms to identify biomarkers and targets that support more rational, durable immunotherapy.
This programme resolves how T-cell clones expand, contract, and redistribute during immunotherapy in small cell lung cancer. Using longitudinal peripheral immune profiling we integrate bulk and single-cell T-cell receptor (TCR) sequencing, single cell (sc) RNA-sequencing and CyTOF (Cytometry by Time-of-Flight) to track clonal behaviour during chemo- and immunotherapy. Antigen-specificity grouping (GIANA – Geometric Isometry-based TCR AligNment Algorithim) allows us to follow functionally related T-cell families across compartments and distinguish the clonal signatures of sustained responders from those of rapid contraction and non-response, informing biomarkers of durable benefit and candidate combination strategies.
This programme pursues two complementary strands in multiple myeloma. The first dissects response and resistance to T-cell engaging therapies (bispecifics, CAR-T [Chimeric Antigen Receptors]), examining how antigen density, tumour burden, antigen escape and the local tumour microenvironment shape efficacy and durability; we track these variables through longitudinal bone marrow and peripehral blood sampling using multi-parameter spectral flow cytometry, Enzyme-Linked Immunosorbent Assay (ELISA), and TCR-sequencing. The second exploits Endoplasmic Reticulum (ER) homeostasis as a source of selective vulnerability in myeloma plasma cells. Our goal is to develop biomarkers for existing myeloma therapies as well as validate therapeutic targets that will the inspire the next generation of myeloma drugs.
Modern immuno-oncology generates biomarker data across a widening set of modalities – bulk and single-cell transcriptomics, TCR repertoires, spectral flow and mass cytometry, spatial and structural biology, imaging, and longitudinal clinical readouts – but the clinical decisions these data are meant to inform still rest on a small number of markers interpreted in isolation. We are developing AI software platforms that integrate multimodal biomarker data to support rational, patient-specific treatment decisions. Drawing on a governance architecture involving multiple integrated AI decision makers our systems are designed to reason across data types, surface the evidence behind each recommendation, and flag uncertainty rather than obscure it. The aim is not to replace clinical and scientific judgement, but to extend it, giving clinicians and researchers transparent, auditable AI collaborators that can keep pace with the scale and complexity of modern biomarker data.
