High-speed and high-ratio referential genome compression.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• High-throughput sequencing generates vast amounts of genomic data, posing storage and communication challenges.
• Traditional compression methods struggle with DNA sequence characteristics like small alphabets and repeats.
• Reference-based lossless compression offers a promising solution for efficient genome data management.

Purpose of the Study:

• To develop a high-performance reference-based genome compression algorithm.
• To address the limitations of existing methods in compressing large genomic datasets.
• To improve the efficiency of genomic data storage, compression, and communication.

Main Methods:

• Developed HiRGC, a novel referential genome compression algorithm.
• Employed a 2-bit encoding scheme and a greedy-matching search on a hash table.
• Compared HiRGC against four state-of-the-art compression methods using benchmark datasets.

Main Results:

• HiRGC achieved compression ratios of 217 to 82 times for human genomes, compressing 21 GB to 96-260 MB in under 30 minutes.
• Performance was at least 1.9 times better and 2.9 times faster than competing algorithms.
• Demonstrated consistent and excellent performance across diverse human and other species' genomes, outperforming methods sensitive to reference genome variations.

Conclusions:

• HiRGC offers a significant advancement in genome compression technology.
• The algorithm provides a stable, robust, and highly efficient solution for managing large-scale genomic data.
• Freely available C++ and Java source codes facilitate adoption in academic and non-commercial research.