Pair-end reads are merged into single file: need to separate

Question

[Chirag Nepal]

Hi there,

I downloaded publicly available datasets. which is 100bp long pair-end reads (with 200 nt of insert size). I downloaded fastq to map to the genome, but it seems authors have merged the reads. I blat few examples and only 100 reads map in stretch.

Fastq example

@SRR893106.1 1 length=202
CATAGGGTGCTCCGGCTCCAGCGTCTCGCAATGCTATCGCGTGCACACCCCCCAGACGAAAATACCAAATGCATGGAGAGCTCCCGTGAGTGGTTAATAGGGGGAGCCTATCATATATCTCCCTACCAACAAACCTACCCACCCTTAACAGCACATAGTACATAAAGCCATTTACCGTACATAGCACATTACAGTCAAATCC
+SRR893106.1 1 length=202
@@@FF?B:CFHHHIJGIIIIIEIHGGIIIJGGGGGGDABBB8;;CDAEHHFFD:?9?BBBDB>ACA:CD:>CCDDBC<(8?>:@?B8>@?:A@ABC3>3@>?<**ACACCAC;::>:;>5
@SRR893106.2 2 length=202
AGACAGATACTGCGACATAGGGTGCTCCGGCTCCAGCGTCTCGCAATGCTATCGCGTGCACACCCCCCAGACGAAAATACCAAATGCATGGAGAGCTCCCGGGGGTAGCTAAAGTGAACTGTATCCGACATCTGGTTCCTACTTCAGGGCCATAAAGCCTAAATAGCCCACACGTTCCCCTTAAATAAGACATCACGATGGA
+SRR893106.2 2 length=202
CCCFFFFFGHHHHJJJJIJJJIAHHEIIIJJIJJJJGIDDEGAA@GGGIIHHHEFF**BBDCCCDDCDCDCDEDDCBCDACDDDD@@CFBDDFHHGHHCGHHIJJJJJJJJIJIJJJJJJIIIIJJJJJJJJJIJJIJJJIJJIJJJJJJJJIHHHHF?@ECECEDCDDEDDDDDCCCDDDBD?@?

I checked sequence using FASTQC which suggest authors have merged reads.

https://www.dropbox.com/s/clpmi7cktr2w7l1/Screen%20Shot%202014-09-18%20at%2017.35.31.png?dl=0

Is there any existing tools or suggest how to separate it.

Thanks in advance !

Cheers

[Devon Ryan]

Did you download the sra file and then forget to use the --split-3 option?

[matted]

Looks like. Fastq files are available for each end separately at the ENA: http://www.ebi.ac.uk/ena/data/view/SRR893106

[Chirag Nepal]

Thanks matted !

ENA has the correct file format, which can be found here.

[Chirag Nepal]

This is what i used to download SRA:

wget http://ftp-trace.ncbi.nlm.nih.gov/sra/sra-instant/reads/ByStudy/sra/SRP/SRP025/SRP025150/SRR893106/SRR893106.sra
wget http://ftp-trace.ncbi.nlm.nih.gov/sra/sra-instant/reads/ByStudy/sra/SRP/SRP025/SRP025150/SRR894447/SRR894447.sra

And default parameter of fastq-dump

for name in $(ls *sra)
do
    echo $name
    ~/unixTools/sratoolkit.2.3.5-2-ubuntu64/bin/fastq-dump $name
done

--split-3 is option on which tool? fastq-dump?

[Devon Ryan]

fastq-dump --split-3 $name

If you don't do that you'll get merged reads like this. Anyway, as matted said, it's usually easier to see if ENA has the fastq files first.