AIMS

AIMS assigns intrinsic molecular subtypes of breast cancer from individual gene expression profiles, providing dataset-independent absolute subtype classification for molecular characterization and prognostic stratification.

Key Features:

• Absolute Subtype Assignment: Assigns one of five intrinsic molecular subtypes—Luminal A, Luminal B, Her2-enriched, Basal-like, and Normal-like—to individual breast cancer samples based solely on gene expression profiles.
• Dataset Independence: Provides absolute subtyping that does not depend on dataset composition or relative comparisons within a cohort.
• Robustness and Stability: Demonstrates high agreement with PAM50 (76% in cross-validation and 77% in the test set) and preserves prognostic value while showing stability across evaluated conditions.
• Versatility Across Technologies: Applicable to multiple gene expression technologies, explicitly including RNA-seq.

Scientific Applications:

• Breast cancer molecular characterization: Enables consistent intrinsic subtype classification for studies of tumor biology and heterogeneity.
• Prognostic stratification: Supports preservation and assessment of prognostic value associated with intrinsic subtypes.
• Cross-study and cross-technology comparison: Facilitates reproducible subtype assignment across datasets and sequencing technologies for comparative analyses.
• Translational research and therapy development: Informs research on targeted therapies and clinical research that rely on subtype-specific molecular information.

Methodology:

Developed by analyzing gene expression data from 4,924 breast cancer patients to define an algorithm that assigns absolute subtypes from individual gene expression profiles, with evaluation using cross-validation, a test set, and benchmark tests including subtype reproducibility between technologies, gene removal, and normal gene expression contamination.