A simple program for getting k-mer counts from a fastq/fasta file, written in Rust.
This command line program takes a fastq/fasta file as input and outputs the counts of k-mers of a specified length. It is implemented using hash tables and a simple algortihm but is still reasonably fast (mostly by using parallel computation with the Rayon library). It can also be used to get per cycle base content for Illumina reads, by setting the k-mer size to the cycle count.
I provide precompiled binaries for linux only here, but it is simple to compile and run:
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh
git clone https://github.com/angelovangel/fastkmers.git
cd fastkmers
cargo build --release
The executable file fastkmers is now under ./target/release/
# Make sure the executable is in your path
# check available options
fastkmers -h
# to get 4-mer counts and a summary
fastkmers -k 4 -s file.fastq.gz
# output json, input fasta
fastkmers -k 4 -j file.fasta
# stdin can also be used as input, use -
cat file.fasta | fastkmers -k 4 -j -
# query for a specific k-mer
fastkmers -k 5 -q "AATTG" file.fastq.gz
# query with regex is also supported
# this example would match all 5-mers whose last 4 bases are: not T| A | T or G | A
fastkmers -k 5 -q "[^T]A[T|G]A$" file.fastq.gz
# get base contents per cycle (the number of cycles has to be known beforehand)
fastkmers -k 126 -c tests/test.fasta
The k-mer counts are printed to stdout as a tab-separated table or as json.
I haven't compared to other programs (e.g. jellyfish), below are some measurements of the execution times for different k-mer sizes of the E. coli MG1655 genome, performed on a MacBook Pro 2018 (Intel i5 and 8 Gb RAM).
hyperfine -r 4 --warmup 1 --export-csv hyperfine-kmer-size.csv -P kmer 4 29 'fastkmers -k {kmer} -a mg1655.fasta'