Soutenances de thèses :

Le 16 avril 2025 à 13:30 - Bat. 9 - salle 109

Présentée par Chen Junyi - IMAG

Modelling, analysis and numerical simulation of the coagulation cascade

Coagulation cascade is a series of biochemical processes critical for preventing bleeding and ensuring wound healing of which disorders lead to pathological issues. However, the existing coagulation models, often lacking experimental validation and simplification, present challenges in their practical application for understanding diseases such as thrombosis and haemophilia. This thesis focuses on mathematical modeling for the coagulation cascade, including optimizing parameters, model reduction, and finally application of venous thrombosis.
First, a novel optimizer of clotting factors using gradient-based and evolutionary optimization algorithms has been developed and used to fine-tune thrombin coagulation models. It is found that the types of models and selection of variables influence the complexity and the landscape significantly affects the optimization. The novel optimizer improves the efficiency to find the global minimum, especially for a very complex landscape.
An other achievement in this thesis is a multi-step reduction approach, simplifying large-scale coagulation models without compromising their predictive capability. This reduction approach combines the advantages of different reduction methods and reduces the size of coagulation models efficiently, while conserving the initial accuracy and robustness.
Finally, this thesis presents a 0-D coagulation model for venous thrombosis. This model incorporates time-dependent transport effects derived from mechanical features of 3-D simulations, along with biochemical interactions from the coagulation cascade. Simulations were conducted for both platelet-based and platelet-free models. The results highlight the critical influence of cascade stimulation, transport effects, and platelet involvement in the development of models for deep venous thrombosis.