Randomizing reads for Salmon

Question

[majd.abdulghani]

Hi Biostars community,

I want to run Salmon with aligned samples. I only have sorted BAMs, and the way Salmon works, it needs the reads to be randomized in the input file. Will sorting by read name effectively shuffle the reads? I have paired-end data, so I know the pairs will stay adjacent when I sort by name. Is that okay for Salmon, or do I need to make sure even the read pairs are shuffled?

Thanks!

[biofalconch]

Will sorting by read name effectively shuffle the reads?

Probably it will, I would suggest having a look at samtools collate first ;)

Is that okay for Salmon, or do I need to make sure even the read pairs are shuffled?

No, read pairs should stay together :)

[majd.abdulghani]

Thank you. Reminders to check the documentation are always helpful. I double checked and am fairly sure it does what I'm expecting it to. I wasn't sure if having read pairs together is or isn't desirable in Salmon's case, and you've very kindly answered that :)

Thanks again!