DNA Repair and Uveal Melanoma (DRUM)

We investigate genetic predisposition to uveal melanoma and the events that drive the transformation of normal uveal melanocytes. By combining primary and metastatic tumour analyses, multi-omic data and genetically defined models, we seek to understand how oncogenic alterations and tumour suppressor loss shape disease evolution. We aim to incorporate machine learning tools to connect molecular trajectories with clinical features.

We investigate determinants of response and resistance to uveal melanoma treatments, including tebentafusp and PKC inhibitors. Our work explores PKC–MAPK–MET signalling, darovasertib–crizotinib combinations, FAK, mechanobiology and the tumour microenvironment. Cellular and three-dimensional models support mechanistic studies and combination testing. We also develop vaccine strategies and seek to integrate clinical, molecular and imaging data, including through machine learning, to guide preclinical validation and early-phase clinical research.

We are developing a programme focused on DNA methylation in uveal melanoma and cancers with DNA repair defects. We seek to understand how epigenetic states contribute to tumour identity, progression and treatment response. Integrating methylome, genomic and transcriptomic data aims to identify biological signatures and biomarkers. Machine learning tools will be evaluated to distinguish tumour subgroups and characterise their heterogeneity.

Building on our discoveries involving MBD4 in uveal melanoma, we study the repair of lesions arising from 5-methylcytosine deamination and the consequences of repair failure. We investigate interactions between MBD4 and the mismatch repair machinery, including non-canonical functions involving MLH1 and PMS1. Human genetics, mutational signature analysis and functional models connect repair mechanisms with cancer predisposition and therapeutic vulnerabilities. Computational approaches support the interpretation of these signatures.

Homologous recombination deficiency is a major research focus, particularly in breast and ovarian cancers. We develop and evaluate genomic signatures and HRD detection approaches, including shallow whole-genome sequencing and optical genome mapping. We investigate their relationships with DNA repair gene alterations, methylation and treatment outcomes. Integrating molecular, histological and clinical data, including through machine learning, aims to refine HRD characterisation and the assessment of biomarkers of treatment sensitivity and resistance.

Our team is led by Manuel Rodrigues, a medical oncologist specialising in ocular and gynaecological cancers, Head of Medical Gynaecology at Institut Curie. His patient care, clinical trial coordination and research activities connect hospital-based questions directly with laboratory programmes. This integration allows us to confront research hypotheses with clinical observations and develop translational projects based on patient samples and clinical data. Manuel leads Institut Curie’s Uveal Melanoma Medical and Scientific Group (BRIO) and co-leads the Gynecological Oncology BRIO.

Develop genetically defined models to understand uveal melanoma initiation and test the functional consequences of oncogenic alterations and BAP1 loss.

Establish a DNA methylation programme on constitutional and tumor DNA.

Dissect BER–MMR interactions and the consequences of repair defects for mutagenesis.

Evaluate HRD signatures and detection methods in breast and ovarian cancers and investigate their associations with treatment response.