My research focuses on systems and cancer biology through mechanistic and predictive modelling, combined with AI/ML methods applied to multi-omics data. I am particularly interested in:

Elucidating mechanisms of adaptive resistance to anticancer therapies 

One of the most formidable obstacles in modern oncology is the development of drug resistance, which remains a primary driver of treatment failure and cancer-related mortality. While acquired resistance, which arises from the selection of pre-existing or newly acquired genetic mutations, has been studied extensively, a more immediate and dynamic challenge is adaptive resistance. This phenomenon is characterized by a rapid, non-genetic rewiring of cancer cell signalling networks that can occur within hours or days of initiating targeted therapy. Mechanisms such as feedback loop disruption, pathway crosstalk, and signalling rebound allow cancer cells to dynamically reprogram their internal circuitry to survive therapeutic assault, rendering initially effective drugs obsolete.

My work has consistently provided the mechanistic insights necessary to understand and ultimately combat adaptive resistance. My approach leverages mechanistic and computational modelling framework to move beyond mere description, revealing the non-intuitive, systems-level consequences of therapeutic intervention.

• Systems modelling of the EGFR-PYK2-c-Met interaction network predicts and prioritizes synergistic drug combinations for triple-negative breast cancer, PLoS computational biology 14 (6), e1006192; https://doi.org/10.1371/journal.pcbi.1006192
• The crossregulation between ERK and PI3K signaling pathways determines the tumoricidal efficacy of MEK, Journal of molecular cell biology 4 (3), 153-163; https://doi.org/10.1093/jmcb/mjs021
• Integrative modeling uncovers p21-driven drug resistance and prioritizes therapies for PIK3CA-mutant breast cancer, npj Precision Oncology 8 (1), 20; https://doi.org/10.1038/s41698-024-00496-y
• Integrative modelling of signalling network dynamics identifies cell type-selective therapeutic strategies for FGFR4-driven cancers, Cancer Research 84 (19), 3296–3309; https://doi.org/10.1158/0008-5472.CAN-23-3409

Decoding complex cellular signalling to reveal fundamental design principles 

At the heart of cellular function lies a vast and intricate network of signalling pathways that process information from the environment and orchestrate appropriate responses, such as proliferation, differentiation, and apoptosis. The behaviour of these networks is not random; it is governed by a set of recurring architectural motifs—often referred to as "design principles"—that have been honed by evolution to enable robust and precise cellular decision-making. These principles include structures like positive and negative feedback loops, which can create switch-like behaviours and oscillations; feed-forward loops, which can filter out transient noise or detect persistent signals; and crosstalk between pathways, which allows for the integration of multiple inputs into a coherent response. In cancer, these exquisitely controlled networks are hijacked and rewired to drive malignant phenotypes.

A central pillar of my research has been a systematic effort to reverse-engineer these design principles within the context of oncogenic signalling, providing a fundamental understanding of how cancer cells compute and make decisions.

• Positive-and negative-feedback regulations coordinate the dynamic behavior of the Ras-Raf-MEK-ERK signal transduction pathway, Journal of cell science 122 (3), 425-435; https://doi.org/10.1242/jcs.036319
• Functional roles of multiple feedback loops in ERK and Wnt signaling pathways that regulate epithelial-mesenchymal transition, Cancer research 70 (17), 6715; https://doi.org/10.1158/0008-5472.CAN-10-1377
• A hidden incoherent switch regulates RCAN1 in the calcineurin–NFAT signaling network, Journal of cell science 124 (1), 82-90; https://doi.org/10.1242/jcs.076034
• The switching role of β-adrenergic receptor signalling in cell survival or death decision of cardiomyocytes, Nature communications 5 (1), 5777, https://doi.org/10.1038/ncomms6777
• A regulated double-negative feedback decodes the temporal gradient of input stimulation in a cell signaling network, Plos one 11 (9), e0162153; https://doi.org/10.1371/journal.pone.0162153
• Coupled feedback regulation of nuclear factor of activated T-cells (NFAT) modulates activation-induced cell death of T cells, Scientific Reports 9 (1), 10637; https://doi.org/10.1038/s41598-019-46592-z

Developing predictive and companion biomarkers to guide treatment

The advent of targeted therapies has revolutionized cancer treatment, but their success is critically dependent on the ability to identify the right patients for the right drug. This is the central role of predictive biomarkers and companion diagnostics (CDx), which are designed to predict the efficacy and/or safety of a specific therapy for an individual patient. The development of trastuzumab for HER2-amplified breast cancer, for example, stands as a landmark success for this paradigm. However, the development of robust and reliable biomarkers remains a significant challenge. Many targeted drugs have limited efficacy because simple, single-analyte biomarkers often fail to capture the complexity of the underlying signalling networks. Furthermore, the increasing use of artificial intelligence and machine learning (AI/ML) in biomarker discovery, while powerful, often results in "black box" models. These models may be predictive, but their lack of mechanistic interpretability is a major barrier to clinical trust and implementation, as clinicians are hesitant to base critical treatment decisions on an algorithm whose reasoning is opaque. 

My research directly confronts these challenges by developing novel computational frameworks that integrate the predictive power of ML with the explanatory power of mechanistic models.

• SynDISCO: a mechanistic modeling-based framework for predictive prioritization of synergistic drug combinations targeting cell signalling networks, Computational Modeling of Signaling Networks, 357-381; https://doi.org/10.1007/978-1-0716-3008-2_17
• A Boolean-based machine learning framework identifies predictive biomarkers of HSP90-targeted therapy response in prostate cancer, Frontiers in Molecular Biosciences 10, 1094321; https://doi.org/10.3389/fmolb.2023.1094321

Building patient-specific QSP models that integrate mechanistic signalling networks to simulate and predict drug responses.

Quantitative Systems Pharmacology (QSP) represents the pinnacle of translational modelling in drug development. It is a sophisticated discipline that integrates multiscale mathematical models with experimental and clinical data to mechanistically connect the pharmacology of a drug (its pharmacokinetics and pharmacodynamics) to the pathophysiology of a disease at the whole-system level. The goal of QSP is to create predictive, in silico models that can simulate clinical trial outcomes, optimize dosing regimens, identify patient populations most likely to respond, and de-risk the entire drug development process. This approach is increasingly endorsed by regulatory agencies like the FDA as a key component of model-informed drug development, recognized for its potential to increase the efficiency and success rate of bringing new therapies to patients.

My body of work constitutes the de facto construction of a powerful and versatile QSP platform for oncology. My studies can be systematically deconstructed and mapped onto the core components and principles of QSP, demonstrating a bottom-up assembly of the essential, validated components required to build patient-specific predictive models. Building a credible QSP model or a "digital twin" is not a single endeavour; it requires a vast library of pre-existing, validated knowledge about the system's components and their interactions. This includes a deep understanding of how signalling pathways function, how drugs perturb them, and what biomarkers define patient heterogeneity. 

• SynDISCO: a mechanistic modeling-based framework for predictive prioritization of synergistic drug combinations targeting cell signalling networks, Computational Modeling of Signaling Networks, 357-381; https://doi.org/10.1007/978-1-0716-3008-2_17
• Systems modelling of the EGFR-PYK2-c-Met interaction network predicts and prioritizes synergistic drug combinations for triple-negative breast cancer, PLoS computational biology 14 (6), e1006192; https://doi.org/10.1371/journal.pcbi.1006192
• Integrative modelling of signalling network dynamics identifies cell type-selective therapeutic strategies for FGFR4-driven cancers, Cancer Research 84 (19), 3296–3309; https://doi.org/10.1158/0008-5472.CAN-23-3409