GeneCompete

GeneCompete prioritizes genes implicated in diseases by integrating multiple gene expression datasets and scoring genes using log-fold change-based ranking methods.

Key Features:

• Union algorithm: Combines results from multiple gene expression datasets into a unified candidate set.
• Ranking methods: Uses Win-Loss, Massey, Colley, Keener, Elo, Markov, PageRank, and Bi-directional PageRank to rank genes.
• Log-fold change scoring: Assigns scores to genes based on log-fold change values between disease and normal samples.
• Winner identification: Identifies genes with significant differential expression as "winners."
• Datasets and technologies: Evaluated on Hypertrophic Cardiomyopathy (HCM) and Microarray Quality Control (MAQC) datasets including microarray and RNA-Sequencing data.
• PageRank plus union: Employs PageRank combined with a union strategy to identify both up-regulated and down-regulated disease-associated genes.
• Validation correlation: Top-ranking genes show strong correlation with TaqMan validation sets across log-fold change thresholds in MAQC datasets.
• Performance benchmarking: Demonstrated superior predictive power compared to classical methods on tested datasets.

Scientific Applications:

• Biomarker discovery: Facilitates discovery of disease biomarkers from microarray and RNA-Sequencing data.
• Gene-disease association: Supports identification and prioritization of genes linked to diseases, including up- and down-regulated candidates.
• Cross-platform integration: Integrates diverse platforms and experimental conditions to improve robustness of gene prioritization.

Methodology:

Applies a union algorithm across multiple gene expression datasets and ranks genes using Win-Loss, Massey, Colley, Keener, Elo, Markov, PageRank, and Bi-directional PageRank on scores derived from log-fold change values, with "winners" defined by significant differential expression.