QVZ

QVZ compresses per-base sequencing quality values using a lossy algorithm to reduce storage of FASTQ and SAM files while preserving genotyping fidelity.

Key Features:

• Lossy compression: Performs lossy compression of sequencing quality values associated with DNA sequencing data.
• File format targets: Operates on quality values embedded in FASTQ and SAM formats.
• Storage reduction: Achieves higher compression ratios than traditional lossless methods, reducing the storage occupied by quality values that typically account for about half of uncompressed sequencing file size.
• Rate–distortion performance: Delivers superior rate–distortion performance across multiple distortion metrics compared to previously proposed algorithms.
• Quasi-convex distortion minimization: Allows minimization of arbitrary quasi-convex distortion functions for customized fidelity criteria.
• Genotyping fidelity: Produces compressed quality values that yield genotyping results closer to those from original quality values at a given compression rate.
• Implementation: Implemented in C.

Scientific Applications:

• Genotyping: Improves fidelity of genotyping analyses when using compressed quality values compared with other compression algorithms.
• Large-scale data storage and transmission: Reduces storage and transmission requirements for large-scale sequencing datasets by compressing quality values.
• Custom downstream analysis: Enables tailoring of quality-value compression to specific downstream analysis requirements via arbitrary quasi-convex distortion functions.

Methodology:

Applies lossy compression with rate–distortion optimization and minimization of quasi-convex distortion functions; implemented in C.