TNBCIS

TNBCIS predicts immune subtypes in triple-negative breast cancer (TNBC) patients to identify candidates for immune checkpoint blockade (ICB) therapy.

Key Features:

• Datasets: Uses gene expression data from The Cancer Genome Atlas (TCGA), Gene Expression Omnibus (GEO), and METABRIC.
• Input data: Operates on gene expression profiles for unsupervised analysis and classification.
• Unsupervised clustering: Performs unsupervised clustering analysis of gene expression profiles to define subtypes.
• Subtype discovery: Identifies two immune subtypes, S1 and S2, within TNBC cohorts.
• Subtype characterization: Characterizes S1 by higher immune scores, elevated levels of immune cells, and a more favorable prognosis relative to S2.
• Comparative analyses: Compares prognostic outcomes, enriched pathways, and ICB indicators between subtypes.
• Validation: Validates subtype associations with overall survival (OS) and relapse-free survival (RFS) using METABRIC samples.
• Hub gene identification: Identifies 11 hub genes: LCK, IL2RG, CD3G, STAT1, CD247, IL2RB, CD3D, IRF1, OAS2, IRF4, and IFNG.
• Predictive modeling: Implements a random forest classifier based on the 11 hub genes achieving an area under the curve (AUC) of 0.76.

Scientific Applications:

• Patient stratification for ICB: Stratifies TNBC patients into subtypes to identify those more likely to benefit from immune checkpoint blockade therapy.
• Prognostic assessment: Associates subtypes with differences in overall survival (OS) and relapse-free survival (RFS).
• Immune microenvironment analysis: Distinguishes immune-infiltration levels and immune scores between subtypes.
• Biomarker discovery: Provides an 11-gene hub set (LCK, IL2RG, CD3G, STAT1, CD247, IL2RB, CD3D, IRF1, OAS2, IRF4, IFNG) for subtype prediction and further study.
• Cross-cohort validation: Enables validation of subtype associations across TCGA, GEO, and METABRIC datasets.

Methodology:

Unsupervised clustering of gene expression profiles from TCGA, GEO, and METABRIC; comparative analyses of prognostic outcomes, enriched pathways, and ICB indicators between subtypes; identification of 11 hub genes (LCK, IL2RG, CD3G, STAT1, CD247, IL2RB, CD3D, IRF1, OAS2, IRF4, IFNG); and development of a random forest classifier achieving AUC = 0.76 with METABRIC validation.