Research
Research mission
Extracting the evolutionary structure hidden in cancer sequencing data.
Our research develops model-based machine learning for bulk, longitudinal, single-cell, spatial and long-read assays, connecting experimental design, uncertainty-aware inference and biological validation.
Technical programme
Methods shaped by the data, the biology and the question.
We develop models for the sequencing data used across modern cancer research. Our work combines statistical and artificial intelligence methods with evolutionary reasoning, allowing us to integrate heterogeneous measurements, quantify uncertainty and recover the latent processes that generate an observed tumour profile. We work from study design and data generation through to computational analysis and biological validation.
01 · MODELS
Machine Learning and Artificial Intelligence
Probabilistic, generative and representation-learning methods designed to remain interpretable, calibrated and biologically meaningful.
02 · DATA
High-throughput Sequencing
Bulk, single-cell, spatial and long-read assays across DNA, RNA and chromatin, including data generated within our collaborations.
From experiment to explanation
This is how we work.
We begin with a biological or clinical question, help design the experiment, generate or curate the right measurements, and build models around the resulting data. Predictions then return to experiments and collaborators for validation.
DESIGN
Frame the question.
MEASURE
Generate the data.
INFER
Model the process.
VALIDATE
Test the explanation.
Key collaborators
Science across institutions.
We work with leading research and clinical institutions in Italy and internationally.
