Abstract
Biological responses to disturbances, such as infections and drug treatments, stem from complex and highly interconnected signaling processes that are challenging to interpret from experimental observations alone. In precision medicine, there is a growing demand for computational methods that not only predict outcomes but also offer mechanistic insights into how biological systems react to external perturbations. This thesis investigates the development and application of mechanistic Boolean modeling to study biological responses, support therapeutic decision-making, and contribute to broader discussions on the responsible use of computational techniques in biomedicine.
The research includes four studies on immune biology, cancer systems pharmacology, and educational reflection on precision medicine. The first study expanded a Boolean model of macrophage signaling to assess how SARS-CoV-2 perturbations affect inflammatory and antiviral responses, revealing that such disruptions can enhance pro-inflammatory signaling while impairing interferon responses. The results underscore how dysregulation at the network level may contribute to severe disease outcomes. In the second study, a Boolean model of colorectal cancer signaling was used to simulate drug perturbations and prioritize candidate combinations for experimental validation. Simulation results indicated that synergy can arise when complementary survival pathways are co-targeted, thereby disrupting compensatory signaling and promoting apoptosis.
Building on these disease-specific applications, the third study developed a scalable computational pipeline for the systematic analysis of drug combinations across various cancer contexts. Using condition-specific Boolean models to simulate drug perturbations and quantify resulting signaling responses, the framework enabled predictive assessment of synergistic responses from large-scale drug screening datasets, while also supporting mechanistic interpretation of combination effects. Comparative analyses showed that many responses were strongly influenced by the dominant effect of a single compound, while similar synergistic outcomes could arise through distinct signaling mechanisms across cellular contexts. The results illustrate the potential of mechanistic modeling to advance drug combination research.
The final study examined the broader implications of computational approaches in precision medicine by developing a scenario-based deliberative learning format. Structured discussions with students revealed that conversations initially focused on the technical opportunities of AI-supported precision medicine. However, as the activity progressed, participants began to consider issues of fairness, responsibility, trust, and the allocation of healthcare resources. This study emphasizes that the successful implementation of computational methods in medicine relies not only on their technical performance but also on ethical reflection and societal dialogue.
The studies collectively show that mechanistic Boolean models can serve as valuable explanatory tools, frameworks for generating hypotheses, and scalable methods for analyzing biological changes. This thesis advances the field of computational modeling from context-specific mechanistic analyses towards preclinical and translational applications in precision oncology. It also highlights the importance of integrating computational methods responsibly, with ongoing attention to interpretability, uncertainty, and societal values.