This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Does Order Within Read Pairs In Interleaved Files Matter?

I have now ended up with interleaved paired-end read files where the order of reads is not the same throughout the file, ie. sometimes the forward read is first, sometimes the reverse read is first, but they are otherwise paired in succession.

I have always assumed that all the programs just look at the order of the reads and headers are ignored, since it does not matter which read is forward, which is reverse, so I should be fine with this for downstream applications?

paired-end reads
• 5,836 views • link
updated by swbarnes2 swbarnes2 • written by DoubleDecker DoubleDecker

3 answers

If it's straight from the aligner, typically read #1 is first, regardless of its orientation. Regardless, anything written to process name-sorted SAM or BAM files should be able to deal with that simply by looking at the flags. If a program can't deal with that, I wouldn't trust its output to begin with since it's likely to be doing many things incorrectly.

• 0 views • link
written by Devon Ryan Devon Ryan

Note that I am talking about .fastq files here.

• 0 views • link
written by DoubleDecker DoubleDecker

Oh, it's incredibly unusual to interleave fastq files like that, which is why, without other context, I assumed you were talking about aligned reads. What are you hoping to achieve by interleaving your fastq files?

• 0 views • link
written by Devon Ryan Devon Ryan

A lot of assemblers accept only interleaved reads as input.

• 0 views • link
written by DoubleDecker DoubleDecker

I would definitely reorder my fastq files the same way to avoid potential troubles, which is not that difficult. For instance, you can sort your two files by ID, take a look here: how to efficiently sort a FASTQ file by entry ID?

• 0 views • link
written by Manu Prestat Manu Prestat

Yeah, you are right, does not take long but better sort the reads by their names just in case. I found this bash script very helpful: cat file.fastq | paste - - - - | sort -k1,1 -t " " | tr "\t" "\n" > file_sorted.fastq

• 0 views • link
written by DoubleDecker DoubleDecker

Just to check, you know that while the two reads of a pair should run in opposite directions, you know that most library preps work so that the direction that read 1 goes in is random, right?

So I should think that you would expect an interleaved file to go read 1, read 2, read 1, read 2, etc, read 1 could go in either direction, and read 2 will go in the opposite direction.

• 0 views • link
written by swbarnes2 swbarnes2

Log in to answer this question.

More posts like this