A Comparative Study on the Evaluation of k-mer Indexing in Genome Sequence Compression

Low-cost and faster next-generation sequencing (NGS) technology generates huge sequence data for living organisms in the terabyte range. Storing, transferring, and analyzing these data is a real challenge for researchers. An efficient compression algorithm is the ultimate solution to this challenge. Three benchmark methods are tested on the Amazon Web Services (AWS) virtual cloud platform in this article: High-performance referential genome compression (HiRGC), High-efficiency referential genome compression (SCCG), and Hybrid referential compression method (HRCM). Eight benchmark human genomes, coronavirus genome, and a few additional species in FAST-ALL (FASTA) and Raw formats are used to test these algorithms. The widely-used FASTA format, which is utilized in GenBank, makes data analysis and reading easier for researchers. A very fast k-mer hashing method is used for indexing, which is efficient for pairwise and batch-wise compression. HiRGC offers a good trade-off between compression ratio and time, according to experimental data. The SCCG technique takes longer to compress data but significantly reduces the amount of available space. HRCM does not perform as well for pairwise compression, but it makes significant strides for batch processing. These facts motivate us to propose an improved and efficient compression algorithm in the future.