Request: Add seqtk shuffle command to randomise order of reads

[peterjc]

I have been creating mock community samples using seqtk sample on some single isolate inputs, something like this:

rm -rf tempR1.fastq tempR2.fastq
for sample in A B C; do
    seqtk sample -s 123 input${sample}_R1.fastq.gz 10000 >> tempR1.fastq
    seqtk sample -s 123 input${sample}_R2.fastq.gz 10000 >> tempR2.fastq
done
gzip tempR1.fastq
gzip tempR2.fastq

In this example my combined FASTQ files will have the reads from sample A, then sample B, and finally sample C - and this ordering may introduce biases in the downstream analysis.

What I would like to do is finish with something like this:

seqtk shuffle -s 123 tempR1.fastq | gzip > mixed_R1.fastq.gz
seqtk shuffle -s 123 tempR2.fastq | gzip > mixed_R2.fastq.gz

Here I am assuming -s would set the random number seed as used in seqtk sample to ensure that both R1 and R2 are randomised in the same way, and the output remains nicely paired.

[tseemann]

@peterjc Until this is implemented, you can use seqkit shuffle