fastq2sam - Converting paired-end fastq files with illumina/DNBSEQ-format quality scores into an unmapped sam/bam file while properly handling the RG tags.
split_fastq - Splitting paired-end illumina/DNBSEQ fastq files by read groups.
Linux users can download static binary files from the release page.
The binaries are built automatically by GitHub Actions (see .github/workflows/build.yml) whenever a v* tag is pushed, and are provided for two architectures:
fastq2sam-linux-x86_64.tar.gz(x86_64 / amd64)fastq2sam-linux-aarch64.tar.gz(ARM64 / aarch64)
Each archive contains the fastq2sam and split_fastq executables.
git clone --recurse-submodules https://github.com/shinichinamba/Fastq2Sam.git
# git submodule update --init --recursive
This software depends on the SeqAn3 and Sharg-parser libraries, and the SeqAn3 library requires g++>=11 and the C++20 support.
mkdir build && cd $_
cmake ..
# If your default C++ compiler is not g++>=11, please specify the compiler by e.g., -DCMAKE_CXX_COMPILER=$CONDA_PREFIX/bin/g++.
# (In my Mac OS, I specified /usr/local/bin/g++-13 or /opt/homebrew/bin/g++)
make
make test
If you prefer static build, edit the line 4 of the CMakeLists.txt so that
set (CMAKE_BUILD_TYPE Release)
We note that in Mac, there is no static file for some dependent libraries, so please set CMAKE_BUILD_TYPE as Debug.
The software does not use libbz2, but the SeqAn3 and Sharg-parser libraries link libbz2 by default. You can tell these two libraries to skip linking libbz2 so that you need not to prepare a static library for libbz2. To do so, edit the following config files:
- seqan3/seqan3-config.cmake
option (SEQAN3_NO_BZIP2 "Don't use BZip2, even if present." ON)
option (SEQAN3_NO_CEREAL "Don't use cereal, even if present." ON)
- sharg-parser/sharg-config.cmake
option (SHARG_NO_BZIP2 "Don't use BZip2, even if present." ON)
To use cmake in WSL2, see https://stackoverflow.com/questions/62879479/every-call-to-configure-file-fails-on-wsl-configure-file-problem-configuring-fi
Then, in my local environment for example, I ran the following commands:
conda create --name gxx13 python=3.7
conda activate gxx13
conda install -c conda-forge -c anaconda gxx_linux-64 gcc_linux-64 sysroot_linux-64 bzip2 zlib
# export CMAKE_PREFIX_PATH="/usr/lib64/:$CMAKE_PREFIX_PATH" # I wrote this in ~/.bashrc, but it might be unnecessary
- added the option "--max-index-mismatch" (default 1), which merges read groups whose index sequences differ within the allowed number of base mismatches (effective only with "--use-index-sequence-for-read-group")
- added a GitHub Actions workflow that builds and tests static binaries for x86_64 and aarch64, and attaches them to the release on "v*" tags
- added the option "--use-index-sequence-for-read-group"
- improved read group parsing
- added the executable "split_fastq"
- supported the files using irregular characters for zero quality
- added the option "--phred"
- fixed an error which was raised for too short (< 2 characters in default) read names
- supported the DNBSEQ-format read names
- initial release