NanoSpring - Tool for compression of nanopore genomic reads in FASTQ format (gzipped input also supported). Compresses only the read sequences (i.e., ignores quality values and read identifiers). Achieves over 3 times better compression ratios than Gzip (for recent basecaller versions) and fast decompression. The algorithm requires C++11 and g++ compiler and works on Linux (tested on Ubuntu 16.04, 18.04 and 20.04).
v0.2 includes performance improvements. Make sure you pull the latest version from github or conda.
Scientific Reports: https://www.nature.com/articles/s41598-023-29267-8
To install directly from source, follow the instructions in the next section.
NanoSpring is available on conda via the bioconda channel. See this page for installation instructions for conda. Once conda is installed, do the following to install NanoSpring.
conda config --add channels defaults
conda config --add channels bioconda
conda config --add channels conda-forge
conda install nanospringNote that if NanoSpring is installed this way, it should be invoked with the command NanoSpring rather than ./NanoSpring. The bioconda help page shows the commands if you wish to install NanoSpring in an environment.
git clone https://github.com/qm2/NanoSpring.git
The instructions below will create the NanoSpring executable in the build directory inside NanoSpring.
If you plan to build and run NanoSpring on an older architectures, then you might need to remove the flag
-msse4.1from thetarget_compile_optionsin CMakeLists.txt (or use flags based on the target architecture). Other places that might need to be changed include the-mavx2flag used for libbsc compilation (in the same CMakeLists.txt file). Finally you might need to change the COMMAND for MINIMAP2 target in the CMakeLists.txt file tomake sse2only=1 && make cleanfor compiling minimap2 without the SSE4 optimized code.
On Linux with cmake installed and version at least 3.12 (check using cmake --version):
cd NanoSpring
mkdir build
cd build
cmake ..
make -j
On Linux with cmake not installed or with version older than 3.12:
cd NanoSpring
mkdir build
cd build
wget https://cmake.org/files/v3.12/cmake-3.12.4.tar.gz
tar -xzf cmake-3.12.4.tar.gz
cd cmake-3.12.4
./configure
make -j
cd ..
./cmake-3.12.4/bin/cmake ..
make -jTest installation by running the test script (from the build directory):
./../util/test_script.sh
Run the NanoSpring compressor executable with the options below:
Allowed options:
-h [ --help ] produce help message
-c [ --compress ] compress
-d [ --decompress ] decompress
-i [ --input-file ] arg input file name
-o [ --output-file ] arg output file name
-t [ --num-threads ] arg (=20) number of threads (default 20)
-k [ --kmer ] arg (=23) kmer size for the minhash (default 23)
-n [ --num-hash ] arg (=60) number of hash functions for minhash (default
60)
--overlap-sketch-thr arg (=6) the overlap sketch threshold for minhash
(default 6)
--minimap-k arg (=20) kmer size for the minimap2 (default 20)
--minimap-w arg (=50) window size for the minimap2 (default 50)
--max-chain-iter arg (=400) the max number of partial chains during
chaining for minimap2 (default 400)
--edge-thr arg (=4000000) the max number of edges allowed in a consensus
graph (default 4000000)
-w [ --working-dir ] arg (=.) directory to create temporary files (default
current directory)
--decompression-memory arg (=5) attempt to set peak memory usage for
decompression in GB (default 5 GB) by using
disk-based sort for writing reads in the
correct order. This is only approximate and
might have no effect at very low settings or
with large number of threads when another
decompressor stage is the biggest memory
contributor. Very low values might lead to
slight reduction in speed.
Note that the compressed files are tar archives consisting of the different compressed streams, although we recommend using the .NanoSpring extension as in the examples shown below.
This section contains several examples for compression and decompression with various modes and options. The compressed file uses the .NanoSpring extension as a convention. The decompressed file uses the .reads extension as a convention. If installed using conda, use the command NanoSpring instead of ./NanoSpring.
For compressing file.fastq using default 20 threads.
./NanoSpring -c -i file.fastq -o file.NanoSpring
Using 10 threads.
./NanoSpring -c -i file.fastq -o file.NanoSpring -t 10
For compressing file.fastq.gz (gzipped fastq files) using default 20 threads.
./NanoSpring -c -i file.fastq.gz -o file.NanoSpring
Decompressing the file.NanoSpring with default 20 threads to file.reads.
./NanoSpring -d -i file.NanoSpring -o file.reads
Compare decompressed file to reads in original file to verify compression was lossless.
cmp file.reads <(cat file.fastq | sed -n '2~4p')
Decompressing the file.NanoSpring with 10 threads to file.reads.
./NanoSpring -d -i file.NanoSpring -o file.reads -t 10
Compressing with kmer size for minhash to be 20, number of hash functions for minhash to be 64, and the overlap sketch threhold for minhash to be 8.
./NanoSpring -c -i file.fastq -k 20 -n 64 --overlap-sketch-thr 8 -o file.NanoSpring
Compressing with kmer size of minimap2 to be 25 and the window size for the minimap2 to be 100.
./NanoSpring -c -i file.fastq --minimap-k 25 --minimap-w 100 -o file.NanoSpring
Compressing with the max number of partial chains during chaining for minimap2 to be 800.
./NanoSpring -c -i file.fastq --max-chain-iter 800 -o file.NanoSpring
Compressing with the max number of edges allowed in a consensus graph to be 8000000.
./NanoSpring -c -i file.fastq --edge-thr 8000000 -o file.NanoSpring