WeGET: predicting new genes for molecular systems by weighted co-expression

Area of Science:

• Genomics
• Bioinformatics
• Systems Biology

Background:

• Identifying novel genes within complex molecular systems is crucial for understanding biological functions.
• Existing methods may not fully leverage the vast amount of available gene expression data.

Purpose of the Study:

• To develop and present the Weighted Gene Expression Tool and database (WeGET).
• To predict novel genes based on correlated gene expression within molecular systems.

Main Methods:

• Utilized a compendium of 465 human and 560 murine gene expression datasets.
• Developed a weighting system prioritizing datasets with harmonious gene regulation.
• Calculated weighted co-expression to rank candidate genes and their orthologs.
• Integrated ranks and p-values using rank-order statistics.

Main Results:

• Successfully predicted novel genes co-expressing with Gene Ontology terms and KEGG/Reactome pathways.
• Provided lists of predicted genes, dataset weights, and contributing cell/tissue types.
• Validated performance using 10-fold cross-validation.

Conclusions:

• WeGET is an effective tool for predicting novel genes through weighted co-expression analysis.
• The tool allows users to predict genes co-expressing with custom query sets.
• This approach enhances the discovery of genes within molecular systems.