IRIS: Identifying the gene regulatory basis of inter-individual susceptibility to immune-related adverse events

• Applicants: Dr. Brian Clarke; Dr. Thomas Walle
• Subject Area: Human Genetics ; Bioinformatics and Theoretical Biology
• Term: since 2026
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 581225477

Deciphering how genetic variation gives rise to disease through altered gene regulation remains a central challenge in biomedical research. Although individual genetic variants influencing gene expression can be identified through expression quantitative trait loci (eQTL) analyses, genes operate within interconnected regulatory systems. Understanding the molecular mechanisms of disease and therapeutic response therefore requires a perspective based on gene regulatory networks (GRNs) that capture causal relationships among genes. Conventional GRN inference methods are constrained by their reliance on observational data, which cannot disentangle correlation from causation. Recent advances in single-cell technologies, however, enable causal inference at cellular resolution. Two complementary approaches are particularly powerful in this regard. Large-scale single-cell eQTL mapping links naturally occurring genetic variation to gene expression, while Perturb-seq —CRISPR-based perturbations coupled with single-cell RNA sequencing— allows direct experimental testing of gene–gene interactions. Yet no existing framework integrates these two modalities to construct causal GRNs. The proposed project, IRIS, will establish a novel computational approach for causal GRN inference that jointly models eQTL and Perturb-seq data. The study focuses on CD4 T cells, which are central regulators of immune responses and key mediators of immune-related adverse events (irAEs)—severe autoimmune reactions induced by cancer immune checkpoint therapies (ICTs). The molecular underpinnings of irAEs, as well as the mechanisms conferring resistance to their immunosuppressive treatment with glucocorticoids, remain poorly understood due to the lack of suitable model systems and integrated analytical tools. IRIS will leverage paired genotype and gene expression data from over 50,000 individuals, alongside the first genome-scale Perturb-seq dataset in primary human CD4 T cells, to infer causal GRNs linking genetic variation to cellular phenotypes. These networks will be employed to identify novel regulators of irAE pathogenesis and determinants of glucocorticoid responsiveness. Key predictions will be validated experimentally through targeted Perturb-seq in CD4 T cells derived from irAE patients exposed to glucocorticoids. The applicant team uniquely combines complementary expertise in machine learning–based GRN inference from Perturb-seq data, large-scale eQTL mapping, and leadership in population-scale single-cell consortia as well as prospective irAE clinical cohorts. By integrating population genetics, single-cell perturbation data, and clinical observations, IRIS will provide a new framework for studying the causal genomic mechanisms underlying immune-related diseases. The resulting methodological advances will be broadly applicable to complex human disorders and contribute to the long-term goal of personalized therapies.

• DFG Programme: Research Grants
• International Connection: Netherlands
• Partner Organisation: Nederlandse Organisatie voor Wetenschappelijk Onderzoek (NWO)
• Cooperation Partner: Marc Jan Bonder, Ph.D.