FaStore

Overview

FaStore is a high-performance short FASTQ sequencing reads compressor.

The compression process happens over multiple steps and the compressor currently consists of 3 tools:

• fastore_bin - distribute DNA reads into bins
• fastore_rebin - re-distribute the DNA reads into bins and clusterize further
• fastore_pack - compress the DNA reads stored in bins

However, for ease of use, automated scripts are provided in the scripts directory to perform compression and decompression.

Usage

FaStore offers a variety of different compression configurations. In order to simplify selection, we created 4 profiles, namely lossless, reduced, lossy and max. To perform automatic compression and decompression, a pair of scripts fastore_compress.sh and fastore_decompress.sh is provided.

For example, to compress a pair of FASTQ files IN_1.fastq and IN_2.fastq in the lossless mode with reads represented in pared-end mode and using 8 processing threads type:

./fastore_compress.sh --lossless --in IN_1.fastq --pair IN_2.fastq --out COMP --threads 8

the compressed files will be stored as COMP.cmeta and COMP.cdata files.

To decompress the archives generated with any of the above mentioned profiles and using 8 processing threads type:

./fastore_decompress.sh --in COMP --out OUT_1.fastq --pair OUT_2.fastq --threads 8

the decompressed files will be stored as OUT_1.fastq and OUT_2.fastq files.

Building

Prerequisites

FaStore currently provides Makefiles for building on Linux and Mac OSX platforms. However, it should also be able to be compiled on Windows platform.

The only prerequisite is the zlib library.

Compiling

FaStore binaries are by default compiled using g++ version >= 4.8 (reguired for C++11 threading support). To compile FaStore using g++ >= 4.8 with C++11 standard and dynamic linking, use the default Makefile file and in the main directory type:

make

Alternatively, to compile using clang, invoke make using Makefile.clang file:

make -f Makefile.clang

The resulting fastore_bin, fastore_rebin and fastore_pack binaries will be placed in bin subdirectory.

However, to compile each subprogram separately, use the makefile files provided in each of subprograms directory.

Citing

Roguski, L., Ochoa, I., Hernaez, M., Deorowicz, S. (2018) FaStore: a space-saving solution for raw sequencing data. Bioinformatics 34, 2748–2756.