Grouped False-Discovery Rate for Removing the Gene-Set-Level Bias of RNA-seq.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• RNA-sequencing (RNA-seq) is a powerful tool for gene expression analysis, often replacing microarrays.
• RNA-seq experiments exhibit a gene-level bias where read counts correlate with transcript length, not solely expression levels.
• This gene-level bias introduces a gene-set-level bias in enrichment analysis, favoring gene sets with longer genes.

Purpose of the Study:

• To address the transcript length bias in RNA-seq gene-set analysis.
• To develop a method for accurately calculating statistical significance in RNA-seq gene-set enrichment.
• To remove gene-set-level bias inherent in RNA-seq data.

Main Methods:

• Introduced FDRseq, a novel gene-set analysis method for RNA-seq data.
• Utilized grouped false-discovery rate to accurately calculate gene-set enrichment scores.
• Developed an R program implementation of the FDRseq method.

Main Results:

• FDRseq effectively controls for transcript length bias in RNA-seq.
• Numerical examples demonstrate the appropriateness of FDRseq for RNA-seq gene-set analysis.
• The developed R program provides a practical tool for researchers.

Conclusions:

• Transcript length bias significantly impacts RNA-seq gene-set analysis.
• FDRseq offers a robust solution for accurate gene-set enrichment analysis in RNA-seq.
• Accurate statistical significance calculation is crucial for reliable biological interpretation of RNA-seq data.