Subdyquency

Subdyquency identifies driver genes in cancer by applying a random walk on a bipartite graph that integrates subcellular localization and variation frequency from genomic data generated by high-throughput sequencing technologies.

Key Features:

• Integration of Biological Properties: Subdyquency integrates subcellular localization and variation frequency into its analysis to account for biological context of genes.
• Bipartite Graph Representation: The method models interactions between genes and their dysregulated states using a bipartite graph.
• Random Walk Algorithm: A random walk algorithm traverses the bipartite graph to assess each gene's influence on dysregulated genes within significant cellular compartments.
• Improved Prediction Metrics: When applied to lung, prostate, and breast cancer genomic data, Subdyquency increased precision, recall, and F-score and prioritized both known and rare candidate driver genes compared to existing methods.
• Focus on Variation Frequency and Impact: Results indicate driver genes typically exhibit higher variation frequencies and substantial impacts on dysregulated genes within cellular compartments.

Scientific Applications:

• Driver gene identification: Identification and prioritization of driver genes in lung, prostate, and breast cancer using genomic datasets.
• Tumor biology characterization: Linking genomic alterations to dysregulated genes and cellular compartments to elucidate tumor biology.
• Novel target discovery: Prioritization of rare or novel driver genes as candidate therapeutic targets.
• Therapeutic development and precision medicine: Providing evidence to support development of targeted therapies and personalized medicine strategies.

Methodology:

Subdyquency constructs a bipartite graph integrating subcellular localization, variation frequency, and gene interactions, then applies a random walk algorithm to traverse the graph and score genes by their connectivity and influence on dysregulated genes.