COAC

COAC performs localized gene co-expression network analysis of single-cell RNA sequencing (scRNA-seq) profiles to extract overlapping gene subnetworks that link cell subpopulations to clinical outcomes and drug responses.

Key Features:

• Localized Gene Co-Expression Network Analysis: Decomposes scRNA-seq profile matrices to infer gene subnetworks representing distinct co-expression patterns within specific cell subpopulations.
• High Accuracy in Cell Type Identification: Identifies cell types by analyzing single-cell gene subnetworks with reported accuracy of 83% across multiple time points within cell phases.
• Correlation with Clinical Outcomes: COAC-inferred subnetworks from melanoma scRNA-seq profiles show strong correlation with survival data from The Cancer Genome Atlas (TCGA).
• Pharmacogenomics Biomarker Prediction: Identifies localized gene subnetworks as pharmacogenomics biomarkers for predicting drug responses with area under the ROC curve between 0.728 and 0.783 using Genomics of Drug Sensitivity in Cancer (GDSC) cell line data.

Scientific Applications:

• Network-based diagnostic biomarker discovery: Extraction of gene subnetworks from scRNA-seq data to identify diagnostic signatures linked to cell-type–specific functions.
• Pharmacogenomics and drug-response prediction: Use of COAC-inferred subnetworks to predict patient- or sample-specific drug responses based on GDSC-derived performance metrics.
• Oncology and personalized medicine research: Linking single-cell co-expression patterns to clinical outcomes, enabling applications in cancer prognosis and therapeutic development.

Methodology:

COAC applies matrix decomposition to scRNA-seq profile matrices and identifies overlapping attribute clusters to extract gene co-expression subnetworks and co-expression patterns.