How to "chop" a large FASTQ PE file in half?

[cement_head]

Hello,

I hesitate to use the following terms: "split", "partition", "trim" - because they all have special connotations.

What I'd like to do is to take a large FASTQ file of PE reads and cut the file into two approximately equal halves. However, I want to do it in such a manner that a given file does not have only one half of a pair of a PE read group. In other words, I want to insure that when the file is halved, or split, that no PE reads are separated.

For example, if I have a file with 11 PE reads, I do NOT want 5 reads in one file and 6 reads in another.

Will FASTQ Splitter work for what I want?

- Andor

[GenoMax]

Hello,

For example, if I have a file with 11 PE reads, I do NOT want 5 reads in one file and 6 reads in another.

- Andor

How would you split a PE dataset otherwise? I assume your PE reads are in two separate files (R1/R2) and the split would have to split both files?

[cement_head]

How would you split a PE dataset otherwise? I assume your PE reads are in two separate files (R1/R2) and the split would have to split both files?

No, there are in ONE file, PE reads, RAW FASTQ.

[GenoMax]

So your paired-end reads are interleaved? Why not use "split -n 2" to divide the original into to two parts. If you have an odd number of fastq records then using an explicit "split -l ((n+1)/2*4)" may be better (n = number of fastq records).

[cement_head]

Didn't work, cut the file perfectly in half (FASTQ Splitter.pl)

This from the developer "Currently FASTQ Splitter is not aware of read pairs. One possible solution is to execute FASTQ Splitter by specifying even number of reads for each part. E.g., with "--part-size 200000 --measure count" you'll get 100k pairs per part. You may have to count the reads and calculate the appropriate number of reads per part before that."

I'll try this next.