Better quality score compression through sequence-based quality smoothing.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Next-generation sequencing (NGS) generates vast amounts of data, necessitating efficient storage solutions.
• Quality scores in NGS data contribute significantly to file entropy and present a target for compression.
• Existing tools attempt to smooth quality scores to enhance compressibility without compromising downstream analysis accuracy.

Purpose of the Study:

• To develop a novel tool, YALFF (Yet Another Lossy Fastq Filter), for compressing quality scores in FASTQ files.
• To improve the compressibility of genomic data by smoothing quality values.
• To maintain high precision in SNP calling and genotyping accuracy.

Main Methods:

• Utilizing the FM-Index, a compressed suffix array, for efficient k-mer dictionary storage.
• Implementing an effective smoothing algorithm to reduce quality score entropy.
• Developing YALFF to process FASTQ files on consumer-grade hardware with limited RAM.

Main Results:

• YALFF significantly improves FASTQ file compressibility through quality score smoothing.
• The tool requires minimal computational resources, running effectively on systems with 5.7 GB of free RAM.
• The smoothing algorithm enhances genotyping accuracy in SNP calling pipelines.

Conclusions:

• YALFF offers an efficient method for compressing NGS data by targeting quality scores.
• The tool provides a practical solution for managing large genomic datasets on accessible hardware.
• YALFF demonstrates the potential to improve both data compression and analytical accuracy in genomics.