Disk-based compression of data from genome sequencing.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• High-coverage sequencing data contains substantial redundancy, posing challenges for efficient compression.
• Existing FASTQ compressors struggle to capture overlapping read redundancy within limited memory.
• Disk-based methods, like Burrows-Wheeler transform (BWT), offer improvements but can be further optimized.

Purpose of the Study:

• To develop a novel compression algorithm specifically for sequencing reads (DNA).
• To efficiently handle the redundancy present in high-coverage genomic datasets.
• To improve compression ratios beyond existing state-of-the-art methods.

Main Methods:

• The proposed method utilizes the concept of minimizers for read compression.
• The algorithm is designed to be conceptually simple and easily parallelizable.
• It focuses on capturing redundancy between overlapping sequencing reads.

Main Results:

• Achieved a compression ratio of 0.317 bits per base.
• Successfully compressed a 134.0 Gbp human genome dataset into 5.31 GB.
• Demonstrated superior compression performance compared to previous methods.

Conclusions:

• The overlapping reads compression with minimizers algorithm offers a significant advancement in DNA data compression.
• This method enables efficient storage and handling of large-scale genomic datasets.
• The algorithm's parallelizable nature facilitates its application in large-scale bioinformatics pipelines.