Illumina Paired End FASTQ

kjsalimian · 01-05-2012, 12:31 PM

Hi guys,

I'm very new to this. I have a paired end data set from HiSeq1000. I want to take the first 10,000 or 100,000 reads out of ~40mil reads to use for tests rather than putting the entire 40 mil reads through the tests.

What is the easiest way to generate files of only the first 100,000 reads??

Thanks.

swbarnes2 · 01-05-2012, 01:18 PM

I'm not really familar with HiSeq data foramts, but there are probably some kind of coordinates, maybe by tile. If the data below from the checked answer is correct, you could use a grep to get only the entries with '[sequencer name]_[run]:[lane]:10:'. That's be a fairly random.

Or simpler, head -400000 to get the top 1000k reads?

The ones in the very beginning are probably at the edge of the flow-cell, and will have more bad quality reads. Pulling from the middle might get you more good reads.

http://biostar.stackexchange.com/que...d-naming-tiles

dpryan · 01-05-2012, 01:19 PM

Have a look at this thread. You'll often find that the first batch of reads in a file are crap (at least from some machines), so you're better off randomly selecting them.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
random subset paired-end fastq	dnusol	Bioinformatics	15	04-17-2016 03:36 AM
fastq-dump and paired end reads	moritz	Bioinformatics	3	01-09-2014 02:57 AM
Fastq: Paired end reads and mapping	cedance	Bioinformatics	7	06-18-2011 01:33 PM
how to identify paired end from qseq or fastq	zhaowei	Bioinformatics	1	02-02-2011 01:46 PM
Why are Illumina paired-end SRA datasets made up of 3 FASTQ files?	Bio.X2Y	Illumina/Solexa	9	12-21-2010 12:36 PM

01-05-2012, 12:31 PM	#1
kjsalimian Junior Member Location: USA Join Date: Dec 2011 Posts: 5	Illumina Paired End FASTQ Hi guys, I'm very new to this. I have a paired end data set from HiSeq1000. I want to take the first 10,000 or 100,000 reads out of ~40mil reads to use for tests rather than putting the entire 40 mil reads through the tests. What is the easiest way to generate files of only the first 100,000 reads?? Thanks.

01-05-2012, 01:18 PM	#2
swbarnes2 Senior Member Location: San Diego Join Date: May 2008 Posts: 912	I'm not really familar with HiSeq data foramts, but there are probably some kind of coordinates, maybe by tile. If the data below from the checked answer is correct, you could use a grep to get only the entries with '[sequencer name]_[run]:[lane]:10:'. That's be a fairly random. Or simpler, head -400000 to get the top 1000k reads? The ones in the very beginning are probably at the edge of the flow-cell, and will have more bad quality reads. Pulling from the middle might get you more good reads. http://biostar.stackexchange.com/que...d-naming-tiles

01-05-2012, 01:19 PM	#3
dpryan Devon Ryan Location: Freiburg, Germany Join Date: Jul 2011 Posts: 3,480	Have a look at this thread. You'll often find that the first batch of reads in a file are crap (at least from some machines), so you're better off randomly selecting them.