Abstract
Modern high-throughput sequencing technologies are able to generate DNA sequences at an ever increasing rate. In parallel to the decreasing experimental time and cost necessary to produce DNA sequences, computational requirements for analysis and storage of the sequences are steeply increasing. Compression is a key technology to deal with this challenge. Recently, referential compression schemes, storing only the differences between a to-be-compressed input and a known reference sequence, gained a lot of interest in this field. However, memory requirements of the current algorithms are high and run times often are slow. In this paper, we propose an adaptive, parallel and highly efficient referential sequence compression method which allows fine-tuning of the trade-off between required memory and compression speed. When using 12 MB of memory, our method is for human genomes on-par with the best previous algorithms in terms of compression ratio (400:1) and compression speed. In contrast, it compresses a complete human genome in just 11 seconds when provided with 9 GB of main memory, which is almost three times faster than the best competitor while using less main memory.
Highlights
The development of novel high-throughput DNA sequencing techniques has led to an ever increasing flood of data
In many projects only genomes from one species are considered. This means that projects often deal with hundreds of highly similar genomes; for instance, two randomly selected human genomes are identical to an estimated 99.9%. This observation is exploited by so-called referential compression schemes, which only encode the differences of an input sequence with respect to a pre-selected reference sequence
Since the project contains slightly more than 1000 genomes, we have only extracted the first 1000 genomes named in these Variant Call Format (VCF) files
Summary
The development of novel high-throughput DNA sequencing techniques has led to an ever increasing flood of data. This means that projects often deal with hundreds of highly similar genomes; for instance, two randomly selected human genomes are identical to an estimated 99.9% This observation is exploited by so-called referential compression schemes, which only encode the differences of an input sequence with respect to a pre-selected reference sequence. Using 9 GB of memory usage, our method performs up to three times faster than the best competitor while still needing less main memory Both variants achieve similar compression rates of approximately 400:1 for human DNA. The memory consumption during index creation is limited as follows: at each step of the index generation we have one raw reference block of size at most BS bytes in main memory plus (roughly) 4∗BS bytes for its compressed suffix tree. The input string is traversed from left to right, and depending on the current characters in the input and in the reference block, different subroutines are executed
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
