ITI

ITI integrates large-scale protein-protein interaction data with DNA microarray and gene expression datasets to extract genomic signatures that predict clinical outcomes such as distant metastasis in breast cancer.

Key Features:

• Network-Based Integration: Integrates protein-protein interactions with DNA microarray and gene expression data using a network-based framework for classification of transcription profiles.
• Genomic Signature Extraction: Derives genomic signatures, including estrogen receptor-specific signatures, that predict distant metastasis and patient outcomes in breast cancer.
• Stability and Generalization: Enhances the stability and generalizability of genomic signatures, with reported stability increases of 11–35% across datasets.
• Algorithmic Innovation: Employs a superimposition technique that overlays protein-protein interaction networks onto gene expression datasets to improve predictive precision.
• Performance Metrics: Validation on two compendia showed increased stability (11–35%) and superior accuracy (53–74%) for estrogen receptor-specific signatures compared to previously published methods.

Scientific Applications:

• Cancer Research: Applied to breast cancer interactome and transcriptome data to support investigation of mechanisms and markers associated with distant metastasis.
• Clinical Outcome Prediction: Produces genomic signatures for predicting patient outcomes and stratifying clinical risk in breast cancer cohorts.

Methodology:

Integration of large-scale protein-protein interaction data with DNA microarray and gene expression datasets via a superimposition process; validation performed on multiple compendia.