antiProfiles

antiProfiles identifies stochastic hyper-variability in gene expression and derives genomic anti-profiles for developing cancer diagnostic and prognostic signatures.

Key Features:

• Gene Expression Anti-Profile Development: Identifies stochastic hyper-variability in gene expression across samples and constructs predictive anti-profiles robust to cancer heterogeneity.
• Single-Chip Microarray Normalization: Processes microarray data using single-chip normalization and quality assessment to maintain data fidelity across studies.
• Universal Cancer Signature: Produces a universal cancer anti-profile that discriminates cancerous from normal tissues with ten-fold cross-validation AUC > 0.92 across tissue types.
• Peripheral Blood-Based Diagnostics: Applies anti-profiles to peripheral blood for colon cancer detection with reported AUC = 0.89 in independent studies compared to AUC = 0.81 for a commercial test.

Scientific Applications:

• Cancer Diagnosis and Prognosis: Derives genomic signatures from expression heterogeneity for early cancer diagnosis and prediction of disease progression.
• Non-Invasive Screening: Applies signatures to peripheral blood samples to enable non-invasive cancer screening approaches.
• Broad Applicability Across Cancer Types: Identifies hyper-variable genes across multiple cancer types to support application of anti-profiles across diverse cancers.

Methodology:

Constructs anti-profiles from experimental findings on gene expression variability, applies single-chip microarray normalization and quality assessment, and evaluates signature performance using ten-fold cross-validation to compute AUC.