Flexible expressed region analysis for RNA-seq with derfinder

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• RNA sequencing (RNA-seq) analysis often relies on gene annotations or transcript assembly, which can be incomplete or inaccurate.
• Existing methods for differential expression analysis may not fully capture regulatory elements or novel transcripts outside known gene boundaries.

Purpose of the Study:

• To introduce derfinder, a novel software package for annotation-agnostic differential expression analysis of RNA-seq data.
• To improve the identification of differentially expressed regions (DERs) at single base resolution.
• To provide a flexible and computationally efficient framework for large-scale RNA-seq studies.

Main Methods:

• Implemented a computationally efficient bump-hunting algorithm for identifying DERs genome-wide.
• Developed a flexible statistical modeling framework supporting multi-group and time-course analyses.
• Introduced new data visualizations for expressed region analysis at base resolution.

Main Results:

• derfinder enables the analysis of hundreds of RNA-seq samples at base resolution using R.
• The software identifies differential expression signals outside of known gene boundaries.
• Simulations demonstrate that base-resolution approaches are robust to incomplete annotation and competitive with feature-level methods.

Conclusions:

• derfinder provides a valuable compromise between full transcript reconstruction and feature-level analysis for RNA-seq data.
• The software facilitates the discovery of novel regulatory elements and transcripts.
• derfinder is available as an R package from Bioconductor, promoting accessibility for researchers.