Research

Work in progress

Research

From Genetic Dependencies to Therapeutic Targets

We are deeply engaged in exploring genetic dependencies to identify novel therapeutic targets for cancer treatment. We focus on understanding how tumours exploit specific genes or pathways for survival, growth, and resistance to treatment, leveraging this knowledge to guide drug discovery efforts. By analysing data from genome-wide CRISPR-Cas9 screens in diverse cancer cell lines, we systematically map cancer-specific vulnerabilities and integrate these findings with genomic, transcriptomic, and proteomic data to uncover actionable targets.

Our active participation in the European Cancer Dependency Map initiative is a cornerstone of our efforts. Through this collaboration, we contribute to the generation and curation of large-scale datasets that highlight genetic dependencies across different tumour types. We also develop computational frameworks to prioritise these dependencies as potential drug targets, incorporating multi-omic biomarkers and drug tractability assessments.

Complementing these efforts, our group works to bridge the gap between in vitro models and patient-derived contexts, using data from advanced models such as patient-derived xenografts (PDXs) to design and train machine learning models capable of predicting drug response from multi-omic data. Through this integration of cutting-edge technologies and collaborative research, we aim to address critical challenges in cancer therapy, including overcoming drug resistance and tailoring treatments to individual tumour profiles.

Analysis and Optimisation of CRISPR/Cas9 screens

We design, implement and maintain bioinformatics methods and original tools for the assessment of cancer pre-clinical models, pre-processing, analysis and visualisation of genome-editing screening data, for correcting technology-specific biases in such data, and for the optimisation of single guide RNA libraries for pooled CRISPR-Cas9 screens and other experimental settings.

In silico Drug Repositioning

We are engaged in a broad, disease-agnostic effort to advance in silico drug repositioning through the integration of large-scale biological data and computational modelling. By systematically analysing high-dimensional datasets—ranging from transcriptomics and proteomics to single-cell and spatial omics—we aim to identify therapeutic opportunities for existing compounds outside of their original indications, particularly in areas of high unmet medical need.

At the core of this effort is a data-driven framework that captures the molecular consequences of disease and pharmacological perturbation across diverse biological systems. We use this framework to detect signature similarities, inverse correlations, and context-specific pathway rewiring between disease states and drug-induced cellular responses. This allows us to generate and prioritise hypotheses for drug repurposing based on mechanistic plausibility, tissue specificity, and biological convergence, while minimising reliance on prior disease classifications.

We also explore the use of graph-based approaches, latent representation learning, and causal inference to model drug-disease interactions beyond linear similarity scoring. This enables us to capture higher-order relationships between perturbagens and disease modules, even in sparsely annotated or poorly understood conditions.

Overall, our repositioning strategy is designed to accelerate therapeutic discovery by bridging computational predictions and actionable translational hypotheses. Whenever possible, we prioritise compounds with favourable pharmacokinetics, clinical safety profiles, and existing regulatory approval, thereby facilitating rapid follow-up in preclinical models or early-stage clinical investigation in collaboration with experimental and clinical partners.

Cancer and Ageing

Ageing is the strongest risk factor for cancer, yet the molecular mechanisms linking age-related cellular changes to tumour vulnerabilities remain poorly understood. We develop computational and experimental approaches to dissect how ageing reshapes cellular fitness landscapes, genetic dependencies, and responses to perturbation. By integrating functional genomics, multi-omics profiling, and machine learning, we aim to identify the molecular programs through which ageing influences cancer initiation, progression, and therapeutic response. A particular focus of our work is the development of ageing clocks and predictive models that capture biological age and reveal actionable vulnerabilities emerging during the ageing process.

Externally funded projects

ERC - DepSHOCK

The oncogenic shock model has been introduced to describe the interplay between pro-survival and pro-apoptotic signals that follow an oncogene’s inactivation and lead to cancer cell death. In our recent work, we demonstrated that identifying novel cancer dependencies through systematic computational analyses of CRISPR functional genetic screens and multi-omic data is an excellent means to discover and prioritise new anti-cancer therapeutic targets. In the DepSHOCK project we extend the oncogenic shock concept to any cancer dependency gene.

We call cancer dependency shock (DepSHOCK) the signal dynamics’ interplay that follows the ablation of any gene essential for cancer survival, i.e., cancer dependency gene. Any DepSHOCK triggers the sudden inactivation of pro-survival pathways and activates death-inducing signals, with both events impacting down-streaming transcription factors, leaving a measurable gene expression footprint. The main hypothesis of the project is that each DepSHOCK footprint can be deconvoluted to infer transcriptional markers of signals that emanate from a dependency gene and drive cancer cell survival through the constitutive activity of specific biological pathways. Here, we propose to investigate this hypothesis across tissues, canonical oncogenic addictions, and novel candidate cancer therapeutic targets.

To this aim, we are establishing an experimental/computational platform for the generation and the analysis of unprecedented transcriptional datasets obtained at both bulk and single-cell level upon genetic perturbation of a rationally selected panel of cancer cell lines and therapeutically relevant dependency genes. The overarching goal of the project will be to study the information content of the DepSHOCK footprints and their potential use as computational tools in early anti-cancer drug discovery through new statistical and machine-learning methods delivering mechanistically grounded therapeutic markers and hints for the design of combinatorial therapies.

AIRC - Cancer Drug Sensitivity Recovery Map

We are developing a groundbreaking Cancer Drug Sensitivity Recovery Map to tackle drug resistance in cancer treatment. Using CRISPR screening and original bioinformatics pipelines, we are identifying genetic vulnerabilities in drug-resistant cancer cells across multiple tumour types and widely used chemotherapies. Funded by the Italian Association for Cancer Research (AIRC), this work prioritises targets that can re-sensitise drug-resistant cancer cells, enabling the design of more effective combination therapies. Our findings are being made available through an accessible web platform, driving innovation in cancer treatment and personalized medicine.

Open Targets - Perturbation Catalogue

We spearhead the meta-analysis and harmonisation of CRISPR datasets as part of the Open Targets Perturbation Catalogue project. We focus on aggregating data from diverse CRISPR screens, ensuring consistency across studies, and developing innovative target prioritisation pipelines tailored for cancer, neurodegeneration, and other human disorders. By integrating and analysing these datasets, we also aim to uncover critical insights into gene function and genetic vulnerabilities across diverse disease contexts. This work is pivotal in creating a unified resource for the research community, driving the discovery of impactful therapeutic targets.

Horizon Europe - NeuroCOV

Within the Horizon Europe funded NeuroCOV consortium, we develop computational methods to identify drug repositioning opportunities against the neurological symptoms of NeuroCOVID. Leveraging high-resolution gene expression profiles from NeuroCOVID patient blood samples and integrating them with publicly available datasets of post-perturbational transcriptomic profiles, we aim to pinpoint therapies capable of reversing or preventing these symptoms. Using a combination of established pattern-matching methods and advanced machine learning techniques, we focus on identifying already approved drugs with the potential to address NeuroCOVID’s underlying molecular disruptions.

Collaborating institutes/entities: KUL, FIMM, DZNE, EURICE, HMGU, BGU, UNIMI, VUmc, UmU
Involved team members: Flavio Passante, Francesco Iorio

Hosted Fellowships

AIRC - PostDoctoral Fellowship

Non-genetic mechanisms of drug resistance in breast cancer: role of cancer-associated fibroblasts

By screening 32 breast cancer cell lines with clinical drugs in cancer-associated fibroblast (CAF)-conditioned medium, we identify drug resistance mechanisms influenced by CAF-secreted factors. CRISPR screens then reveal key genes driving this resistance, guiding combinatorial therapies. This AIRC-funded project aims to refine precision treatments by incorporating tumour microenvironment factors, improving patient outcomes.

Collaborating institutes/entities: University of Turin
Hosted Awardee: Aurora Savino

SEMM / Nerviano Medical Sciences - Industry PhD Fellowship

Computational identification and prioritisation of antibody-drug conjugate targets
We aim to advance the development of Antibody-Drug Conjugates (ADCs) by leveraging multi-omic datasets to identify and prioritise cancer-specific surface protein targets. Our approach integrates genomic, transcriptomic, and proteomic data to pinpoint actionable biomarkers, predict ADC efficacy, and refine target-payload combinations. By focusing on cancer heterogeneity and dependency signals, this project aims to enhance the precision, selectivity, and therapeutic potential of ADCs, paving the way for personalised cancer treatments.

Hosted Awardee: Lorenzo Mathieu Brochier

Page updated

Google Sites

Report abuse