Sequence parsing question

Question

[predeus]

Hi all,

I was wondering if anybody knows a smart way (a command or a one-liner) to accomplish the following task.

There is a paried-end fastq file with lots of N's in it.

I need to get a subset of 25,000 matched paired-end reads with no Ns in it.

seqtk seq has a tool that does something similar, but it doesn't work with paired-end reads.

Would welcome any suggestions!

Thank you in advance.

[cpad0112]

try --max-n 0 with cutadapt

Answer 1

[GenoMax]

Untested but should work using reformat.sh from BBMap suite.

reformat.sh -Xmx4g in1=R1.fq.gz in2=R2.fq.gz out=stdout.fq.gz maxns=0 | reformat.sh -Xmx4g in=stdin.fq.gz out1=Sampled.R1.fq.gz out2=Sampled.R2.fq.gz samplereadstarget=25000

If enough reads don't remain after initial filtering you could upsample=t to get 25K.

[predeus]

neat, thank you!