easier

easier estimates patients' likelihood to respond to immune-checkpoint blockade (ICB) therapies by deriving cancer-specific systems biomarkers of anti-cancer immune response from RNA-seq data.

Key Features:

• RNA-seq Data Integration: Integrates RNA-seq data with multiple sources of prior knowledge to extract quantitative descriptors of the tumor microenvironment.
• Immune Repertoire Composition: Quantifies composition of the immune repertoire as one descriptor of the tumor microenvironment.
• Cellular Communication Analysis: Assesses activity of intra- and extra-cellular communications as part of tumor microenvironment characterization.
• Multi-task Machine Learning: Employs multi-task machine learning trained on TCGA data to simultaneously predict several state-of-the-art hallmarks of anti-cancer immune response.
• Cancer-specific Modeling: Derives cancer-specific models and identifies cancer-specific systems biomarkers of immune response.
• Experimental Validation: Biomarkers have been experimentally validated in the literature, and prediction performance has been validated using independent datasets from four different cancer types with patients treated with anti-PD1 or anti-PDL1 therapy.

Scientific Applications:

• ICB Therapy Response Prediction: Predicts patient response to anti-PD1 and anti-PDL1 immune-checkpoint inhibitor therapies using available RNA-seq data.
• Tumor Microenvironment Characterization: Provides holistic description of immune responses in complex and dynamic tumor systems through systems-based signatures.
• Immunotherapy Biomarker Discovery: Identifies interpretable systems biomarkers for immunotherapy efficacy across cancer types.

Methodology:

RNA-seq data is combined with prior knowledge sources to derive system-based signatures quantifying immune-cell composition and intra- and intercellular communications. Multi-task learning is applied to these signatures to predict different hallmarks of immune responses and derive cancer-type-specific models.