Artificial intelligence to investigate metabolomics data for precision medicine

Area of Science:

• Metabolomics and Bioinformatics
• Artificial Intelligence in Healthcare
• Precision Medicine

Background:

• Metabolomic data analysis is key for understanding disease mechanisms and identifying therapeutic targets.
• Traditional methods struggle with high-dimensional, nonlinear metabolomic data.
• AI and ML offer enhanced sensitivity and adaptability for metabolomic data analysis.

Purpose of the Study:

• To analyze and compare AI/ML approaches in metabolomics research.
• To assess the implications of AI/ML in precision medicine.
• To review advancements in AI/ML applications for metabolomic data.

Main Methods:

• Systematic review of peer-reviewed research indexed in PubMed.
• Analysis of AI/ML applications across various diseases like cancer, cardiovascular diseases, and diabetes.
• Comparison of scientific goals, methodologies, datasets, and sources of AI/ML approaches.

Main Results:

• Support Vector Machines (SVM), Random Forests (RF), Gradient Boosting, and Logistic Regression are commonly used AI/ML techniques.
• These methods demonstrate effectiveness in processing complex metabolic data.
• Identified challenges include small cohort sizes, data heterogeneity, and interpretability.

Conclusions:

• AI/ML holds transformative potential for metabolomics.
• AI/ML is critical for advancing precision medicine through predictive metabolic profiling.
• Future work must address current challenges for broader AI/ML adoption in metabolomics.