CRAM 3.1: advances in the CRAM file format

Area of Science:

• Genomics
• Bioinformatics
• Data Compression

Background:

• The Coordinate-Sorted Indexed (CRAM) file format is a widely adopted, high-compression alternative to the Binary Alignment Map (BAM) format for storing DNA sequencing data.
• Continuous development is necessary to optimize data compression techniques for evolving high-throughput sequencing technologies.

Purpose of the Study:

• To describe updates to the CRAM file format (version 3.1) aimed at improving compression efficiency for modern sequencing instruments.
• To quantify the compression gains achieved by CRAM 3.1 compared to previous versions and the BAM format.

Main Methods:

• Implementation of compression algorithm improvements within the HTScodecs library.
• Integration of updated CRAM specifications into the HTSlib library.
• Evaluation of compression ratios using Illumina and long-read sequencing data.

Main Results:

• CRAM version 3.1 achieves 7-15% greater compression than CRAM 3.0 for Illumina sequencing data.
• CRAM 3.1 offers substantial size reductions of 50-70% compared to the equivalent BAM files.
• Compression gains for long-read sequencing data are more modest due to inherent high-entropy signal characteristics.

Conclusions:

• The updated CRAM 3.1 specification provides enhanced data compression for DNA sequencing, particularly for Illumina platforms.
• These improvements lead to significant storage savings, making CRAM a more efficient format for genomic data management.
• The CRAM 3.1 enhancements are readily available through the OpenSource HTScodecs library and HTSlib.