Gene-specific exponent-corrected normalization for library size in bulk RNA-seq

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Bulk RNA-sequencing (RNA-seq) requires library size normalization to distinguish biological variation from technical noise.
• Existing normalization methods often fail to fully remove library size effects in large datasets, impacting downstream analyses.

Purpose of the Study:

• To develop a novel normalization method, gecco, that effectively corrects for residual library size bias in RNA-seq data.
• To improve the accuracy of differential expression and pathway enrichment analyses by mitigating technical noise.

Main Methods:

• Systematic analysis of over 100 public RNA-seq datasets (GEO, TCGA).
• Development of gecco, a gene-specific exponent-corrected normalization method.
• Validation using simulation studies and real-world large-scale RNA-seq datasets.

Main Results:

• Residual library size association observed in standard normalization methods across numerous datasets.
• gecco successfully removes library size bias, yielding normalized counts free of these effects.
• Improved detection accuracy and more biologically relevant pathway enrichment results.

Conclusions:

• gecco offers a robust solution for library size normalization in bulk RNA-seq.
• The method enhances the reliability of differential expression and rhythmicity analyses.
• gecco is implemented in R and compatible with DESeq2 and edgeR.