FastqSubsample

Subsample sequenced reads in FASTQ formt using Julia

Features

• Faster than seqtk when subsampling 100M reads from a FASTQ file with 481M reads
• Smaller 'maximum memory usage' than seqtk when subsampling 100M reads from a FASTQ file with 481M reads

Methods

Reservoir sampling (https://en.wikipedia.org/wiki/Reservoir_sampling) + Loading FASTQ twice

Install

Pkg.clone("git@github.com:yuifu/FastqSubsample.git")

Usage

using FastqSubsample.jl
FastqSubsample(ifastq, ofastq, nSubsample, seed = seed)

• ifastq: File path of an input FASTQ file. Interpret as gzipped file if it ends with .gz.
• ofastq: File path of a subsampled FASTQ file. Interpret as gzipped file if it ends with .gz.
• nSubsample: The number of reads after subsampling.
• seed: Seed. The same seeds will generate the same output files.

For paired end reads:

using FastqSubsample.jl
seed = 123456
FastqSubsample("in.R1.fastq.gz", "out.R1.fastq.gz", nSubsample, seed = seed)
FastqSubsample("in.R2.fastq.gz", "out.R2.fastq.gz", nSubsample, seed = seed)

Docker image

You can run FastqSubsample as a docker image.
Note that you need to specify directories to mount using -v option.

docker run --rm yuifu/fastqsubsample:1.0.0 $ifastq $ofastq $nSubsample $seed

Docker Hub: https://hub.docker.com/r/yuifu/fastqsubsample/