A diagnostic algorithm for inherited metabolic disorders using untargeted metabolomics

Area of Science:

• Biochemistry
• Genetics
• Computational Biology

Background:

• Untargeted metabolomics shows promise for screening rare inherited metabolic disorders (IMDs).
• Challenges include disease rarity, sample scarcity, and incomplete pathway knowledge.
• Current manual diagnostics are slow, and data-driven methods struggle with small sample sizes.

Purpose of the Study:

• Develop a diagnostic algorithm for IMDs.
• Address challenges of small sample sizes and evolving datasets.
• Improve diagnostic accuracy for rare metabolic diseases.

Main Methods:

• Collected 77 IMD patient and 136 control samples.
• Analyzed metabolome using liquid chromatography-mass spectrometry.
• Developed a sparse hierarchical clustering algorithm for IMD signature generation and iterative refinement.

Main Results:

• Algorithm showed iterative improvement, correctly identifying diagnoses within the top 3 in 60% of samples (top 1 in 42%).
• A literature-based case study with 95 IMD samples achieved 73.5% correct diagnosis.
• Demonstrated effective diagnosis even with limited sample availability.

Conclusions:

• The algorithm offers a flexible, data-driven framework for IMD diagnosis.
• Continuous data integration refines diagnostic signatures.
• Enables improved diagnostic capabilities for rare inherited metabolic disorders.