I am having the same issue as this user:

http://user.list.galaxyproject.org/FASTQ-splitter-produced-empty-dataset-please-help-td4654388.html

I have a fastq file, in which both paired end reads were concatenated into a single file. 

@HWI-ST632:288:C2L0BACXX:8:1101:1639:1746 1:N:0:GCCAAT

NTTTGTCTCTGGTCTGTACTTGTGGGCCAGCTTAAGCAGCTGAGTAGCTGTTTGGCGGTCCAGGGCCTGGGTGAACTGGTTAATCGCAGGAGGCACTTTCA

+

#1=DDFFFHHHGFHJIHJIJJIHHJJJIIIIIJJGIGIGIJIGIDGHJIIJJJJJIJJGHHGFFFFDDDDDCBBDCCDDCDDDEDDDDDDBDDDDBDDCCC

@HWI-ST632:288:C2L0BACXX:8:1101:1876:1741 2:N:0:GCCAAT

NTTTATTTTTCGTTATTGTTGGTGGTTTAAAAAATTCCCCCCATGTAATTATTGTGAACACCTTGCTTTGTGGTCACTGTAACATTTGGGGGGCGGGACAG

Fastq splitter results in empty data sets because, as explained in the thread, the data is already split into their corresponding forward and reverse reads and splitter is meant to split reads in which the paired end reads have been joined tail to head. 

I tried the manipulate FASTQ tool as the thread above suggests, however the FASTQ sequence identifier is slightly different so the suggested Match of .1 and .2 example shown in the thread doesn't work, because my FASTQ file has the identifier in Casava 1.8 format, in which the identifier line does not end with the 1 or 2 for forward and reverse strand. Instead the strand is identified as the 1 or 2 after the space in the identifier line, so <space>1:N:0 . . . and <space>2:N:0 . . . .  I tried running FASTQ manipulator with the Match Reads according to regular expression with " 1:" and " 2:" but that did not work, I ended up with two identical fastq files with same sequences as the original file, so it basically just made a copy and did not deconcatenate the forward or reverse. 

Am I using this Match Reads by regular expression wrong?

Janice Patterson

patterja@ohsu.edu