Merging Paired-End FastQ Files

[jmpi]

Hi,

I want to merge the two corresponding FastQ files to get longer reads (with FLASh or COPE).

It seems like I have to sort the FastQ files before.

Does anyone have an idea how I can sort FastQ files by their identifier?

Thank you.

[yangliao]

If you want to sort the reads in every FastQ file by the read ids, I would like to use awk:

# cat my_input.fastq | awk 'BEGIN{line_no=0;} line_no<=3{line_str = line_str $_ "\t"; line_no ++;} line_no == 4{print line_str; line_str="";line_no = 0}'| cut -f 1-4|sort -k 1,1 | sed 's/\t/\n/g' > my_output.fastq

I tested this command on a short FastQ file; it should be OK for longer ones.

[kmcarr]

Hi,

I want to merge the two corresponding FastQ files to get longer reads (with FLASh or COPE).

It seems like I have to sort the FastQ files before.

Does anyone have an idea how I can sort FastQ files by their identifier?

Thank you.

First, what do you mean by "sorted" and are you sure you need to do any kind of sorting? All current sequencers output paired end read data in ordered files. By ordered I mean that the first read in the R1 file matches the first read in the R2 file, second R1 == second R2 and so on. This is what almost all tools which take paired input files expect. The reads do not have to "sorted" in any conventional sense (e.g. alphabetically, numerically) they just have to be in matched order.

Unless something was done to your read files after they were output by the sequencer they will be in matched order and you shouldn't have to do anything with them to input them to FLASH or COPE.

[jmpi]

I just found out that I did not get the original sequence. So, my problem is solved. Thank you both for your help.