WeGet

WeGet predicts novel genes co-expressed with a specified set of query genes by leveraging weighted analysis of extensive human and murine gene expression datasets to identify candidates associated with molecular systems, Gene Ontology terms, and pathways such as KEGG and Reactome.

Key Features:

• Extensive dataset utilization: Employs a compendium of 465 human and 560 murine gene expression datasets derived from various tissues and experimental conditions.
• Weighted co-expression analysis: Assigns higher weights to datasets where known genes of a molecular system show consistent up- or down-regulation to enhance candidate identification.
• Integrated cross-species ranking and statistics: Calculates weighted ranks for human genes and their mouse orthologs and integrates them into a unified gene rank and p-value using a rank-order statistic.
• Pathway and Gene Ontology prediction: Predicts novel genes associated with specific Gene Ontology terms and pathways from KEGG and Reactome.
• Custom query set analysis: Accepts custom query gene sets to generate tailored co-expression-based predictions.
• Performance validation: Assesses predictive performance using 10-fold cross-validation.

Scientific Applications:

• Novel gene discovery for molecular systems: Identifies candidate genes co-expressed with a query set to expand molecular system membership.
• Cross-species genomic inference: Integrates human and murine data and ortholog ranks to support cross-species comparisons and insights.
• Functional annotation of pathways and GO terms: Supports identification of genes associated with KEGG, Reactome, and Gene Ontology annotations.
• Hypothesis generation and validation in genomics: Facilitates exploratory and targeted investigations into gene function and regulation based on co-expression patterns.

Methodology:

Weights datasets where known genes show consistent up- or down-regulation, computes weighted ranks for human genes and mouse orthologs, integrates ranks into a unified gene rank and p-value using a rank-order statistic, and evaluates performance with 10-fold cross-validation.