Implementing metabolic transformation algorithms and their application in ageing-related research

• Autor(a) principal: Sá, Bruno
• Data de Publicação: 2025
• Outros Autores: Oliveira, Alexandre Rafael Machado, Rocha, Miguel
• Idioma: eng
• Título da fonte: Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)
• Texto Completo: https://hdl.handle.net/1822/95197

This work presents the Python-based implementation and validation of the Metabolic Transformation Algorithm (MTA) and its robust counterpart (rMTA), originally developed in MATLAB. MTA/rMTA facilitates the identification of therapeutic metabolic interventions by simulating gene knockouts and evaluating their potential to redirect metabolic fluxes toward healthier phenotypes. The transition to Python capitalizes on its open-source ecosystem, strong scientific computing libraries, and integrated frameworks for constraint-based modeling and optimization. The new Python toolkit utilizes the Gurobi Python API for addressing core Mixed-Integer Quadratic Programming (MIQP) tasks and offers a modular, extensible pipeline for efficiently simulating metabolic perturbations. Validation against the original MATLAB implementations, including studies on RRM1 and RRM2 gene knockouts, confirmed strong consistency in outputs, ensuring methodological reliability and reproducibility. Beyond replication, the pipeline was applied to an ageing-related case study in Caenorhabditis elegans, focusing on metabolic responses to the knockdown of the unc-62 gene. This analysis uncovered key candidate genes and pathways of potential therapeutic relevance, illustrating rMTAs capacity to navigate complex solution spaces and refine intervention strategies. A key innovation of this work is the integration of Evolutionary Algorithms (EAs) into rMTA, enabling the optimization of multi-gene knockout strategies. By exploring synergistic gene deletions, the enhanced approach identifies interventions that more effectively shift metabolic states toward healthier phenotypes. Overall, the Python-based MTA/rMTA software, augmented by evolutionary optimization techniques, provides a more accessible, scalable, and versatile resource for metabolic engineering and therapeutic target discovery. Its successful application to ageing-related metabolic research underlines its value for tackling intricate biological questions and advancing computational approaches in systems biology.