A robust computational pipeline for model-based and data-driven phenotype clustering.

Area of Science:

• Systems Biology
• Computational Biology
• Precision Medicine

Background:

• Precision medicine requires advanced patient stratification methods beyond traditional approaches.
• Complex diseases often present challenges due to limited data and intricate biological processes.
• Standard statistical methods may be insufficient for diseases with poor data availability.

Purpose of the Study:

• To develop an innovative method for phenotype classification.
• To integrate experimental data with mathematical disease models.
• To enable accurate patient stratification for precision medicine.

Main Methods:

• Developed a phenotype classification method combining experimental data and mathematical modeling.
• Utilized mathematical models to infer additional subject features for classification.
• Employed an algorithm to identify optimal clusters and classify samples based on estimated features.

Main Results:

• The methodology proved accurate and robust in both in silico and clinical test cases.
• Successfully classified patients for dyslipidemia (complex disease) and a lysosomal rare disorder.
• Inferred an additional phenotype division not apparent from experimental data alone.

Conclusions:

• The proposed method offers a robust approach for patient stratification in precision medicine.
• This technique enhances understanding of complex diseases, especially with limited data.
• The approach facilitates more tailored medical decisions and treatments.