Machine learning and metabolic modeling-based identification of hypoxia-driven metabolic signatures in pediatric cancers

Key Points

• The aim is to identify metabolic signatures driven by hypoxia in pediatric cancers using machine learning and metabolic modeling.
• Developed pediatric cancer gene expression models (GEMs)
• Utilized scripts for generating parsimonious flux data and machine learning features
• Implemented machine learning analysis and feature interpretation using SHAP
• Included checks for thermodynamically feasible cycles and nutrient uptake rates
• Generated parsimonious flux data reflecting hypoxia-driven metabolic signatures
• Identified machine learning features linked to metabolic processes affected by hypoxia
• Proposed potential pathways for targeting pediatric cancers based on metabolic insights