Causal network perturbations for instance-specific analysis of single cell and disease samples.

Area of Science:

• Genomics
• Systems Biology
• Computational Biology

Background:

• Complex diseases arise from disruptions in multiple biological pathways.
• Characterizing these pathway perturbations in individual patients is crucial for personalized medicine.
• Current methods often overlook data dependencies and pathway limitations.

Purpose of the Study:

• To develop a novel computational approach for subtyping samples based on gene network deregulation.
• To address limitations of existing methods that rely on external databases for pathway analysis.
• To improve the identification of disease mechanisms for enhanced diagnosis and treatment.

Main Methods:

• Introduced sample-specific network perturbation analysis (ssNPA), a method that learns causal graphs from control data.
• Quantified sample-specific network neighborhood deregulation using prediction errors from Markov blankets.
• Evaluated ssNPA on single-cell RNA-seq liver development data and TCGA datasets.

Main Results:

• ssNPA successfully recovered cell timing in liver development data.
• ssNPA identified patient clusters with significant survival differences in TCGA datasets.
• ssNPA consistently outperformed alternative methods across all analyses.

Conclusions:

• Network-based approaches, like ssNPA, offer advantages for analyzing complex diseases.
• ssNPA provides a robust method for subtyping samples based on gene network deregulation.
• This approach facilitates patient-specific identification of disease mechanisms and personalized treatment strategies.