MitoCore: a curated constraint-based model for simulating human central metabolism

Understanding how diseases alter central metabolism remains a challenge due to the complexity of metabolic networks. While genome-scale models have improved, many struggle to simulate abnormal metabolism accurately. These models often include errors in reaction localization, transport steps, and cofactor handling. Large model sizes also hinder parameterization and result interpretation. This gap motivated the need for smaller, manually curated models that better capture the nuances of human metabolism. Prior research has shown that constraint-based models can simulate metabolic behavior under normal conditions. However, no prior work had resolved the inconsistencies in transport steps or cofactor handling. The limitations of existing models highlight the importance of developing more accurate representations of central metabolism. This work addresses the need for a model that integrates mitochondrial and cytosolic metabolism with improved accuracy. The lack of a model that differentiates prosthetic groups from free metabolites also represents a key gap. By focusing on these issues, this study contributes to the field of metabolic modeling.

The aim of this study was to create a more accurate constraint-based model of human central metabolism. The specific problem addressed is the inability of existing models to simulate abnormal metabolism due to errors in reaction localization and transport steps. The motivation stems from the need to better understand how metabolic changes contribute to disease. The study sought to develop a model that could simulate both normal and abnormal physiological conditions. A key objective was to improve the representation of mitochondrial and cytosolic compartmentalization. The researchers also aimed to differentiate prosthetic groups from free metabolites in reactions. Another goal was to enhance the modeling of the respiratory chain and proton motive force. The final objective was to provide a model that is both accurate and easy to interpret.

The researchers developed MitoCore, a manually curated constraint-based model of human central metabolism. The model includes 324 metabolic reactions and 83 transport steps between mitochondrion and cytosol. It also accounts for 74 metabolite inputs and outputs through the plasma membrane. The model incorporates a more accurate partitioning of metabolism between cytosol and mitochondrial matrix. The researchers improved the modeling of connecting transport steps between compartments. They also differentiated prosthetic groups from free co-factors in reactions. A new representation of the respiratory chain and proton motive force was implemented. Finally, the model includes extensive annotations and cross-references to databases like Virtual Metabolic Human and KEGG.

MitoCore successfully simulates human central metabolism under normal and abnormal conditions, including hypoxia and mitochondrial diseases. The model includes 324 reactions, 83 transport steps, and 74 metabolite inputs and outputs. It features improved compartmentalization between cytosol and mitochondrial matrix. The model better represents transport steps between compartments. It also differentiates prosthetic groups from free co-factors in reactions. A new representation of the respiratory chain and proton motive force was implemented. The model's default parameters simulate normal cardiomyocyte metabolism. Over 100 reactions were modified or added compared to Recon 2 to improve accuracy.

The authors propose that MitoCore provides a more accurate and interpretable model of human central metabolism. The model's innovations include improved compartmentalization and transport steps. It also differentiates prosthetic groups from free co-factors in reactions. The new representation of the respiratory chain and proton motive force enhances accuracy. The model's default parameters simulate normal cardiomyocyte metabolism effectively. The extensive annotations and cross-references improve usability and comparison with other models. The researchers suggest that MitoCore can serve as a research tool for experimentalists and modellers. They also propose that it can function as a teaching tool for those new to metabolic modeling.