Boiler: lossy compression of RNA-seq alignments using coverage vectors

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• RNA-sequencing (RNA-seq) generates large datasets, posing storage and computational challenges.
• Existing compression tools often struggle to balance file size reduction with data accessibility for downstream analysis.

Purpose of the Study:

• To introduce Boiler, a novel software tool for efficient compression and querying of large RNA-seq alignment collections.
• To evaluate the impact of Boiler's compression strategy on data recovery and downstream RNA-seq analysis.

Main Methods:

• Boiler compresses RNA-seq alignments by retaining only genomic coverage vectors and empirical distributions, discarding most per-read data.
• The study assessed the storage footprint reduction compared to other compression tools.
• Downstream analysis tools were used to evaluate the impact of Boiler compression on isoform assembly and quantification.

Main Results:

• Boiler achieves a significantly smaller storage footprint compared to existing compression tools.
• Data recovery from Boiler compression shows only a slight negative impact on isoform assembly and quantification.
• Boiler enables fast and efficient querying of compressed alignment data without full decompression.

Conclusions:

• Boiler offers an effective solution for managing large RNA-seq datasets by providing substantial storage reduction.
• The tool facilitates rapid data querying and maintains high utility for critical downstream RNA-seq analyses.
• Boiler is a valuable, free, and open-source resource for the bioinformatics community.