Velvet paired end after some sequences removed?

LizBent · 01-27-2012, 06:23 AM

Hi,

After trimming and filtering my Illumina sequences (paired end 100 bp) for quality, I am left with files that probably don't have the same number of sequences in them any longer. This will be a problem if I try to interleave the files and use them as input for Velvet, right?

Would it be better to:

1. Don't remove any sequences from my files, even if very short or 0 bases, so they can be interleaved
2. Remove sequences and use the (smaller) files in Velvet but not as paired-end reads

Are there other options I'm not aware of?

Thanks

jjohnson · 01-27-2012, 10:22 AM

I would bin the valid pairs and singletons (those with mates removed due to quality trimming/filtering) into 2 separate fastq files. Velvet can accept mutiple files and then you can paramertize around the files (such as specifying insert sizes for mates file, etc).

i.e.

velveth Assem 35 -shortPaired -fasta pe_lib1.fasta -short3 se_lib1.fa

nposnien · 03-05-2012, 01:29 PM

Hi,
we are facing the same problem at the moment. We will have uneven files (one for each pair) after trimming/filtering.

My question is, if there is a script/program out there that would find the mates in in two different files (or in one file if I would merge/shuffle the files prior to trimming/filtering) and bins the unpaired reads into an extra file?

Any help is highly appreciated!

rahularjun86 · 03-05-2012, 01:42 PM

Dear nposnien,
you can use Sickle tool (https://github.com/najoshi/sickle). You only need to input the pair fastq files, and other parameters (scoring system used, quality score to keep and length cutoff etc.), and it will generate the paired and singleton files.
If you want to filter out reads with N's, Just replace the whole sequence with N and quality with #, then set Sickle length and quality values. This way it will filter out reads with N's.
Best wishes,
Rahul

LizBent · 03-06-2012, 12:21 AM

This script may be useful for interleaving pairs for Velvet (and generating non-paired singleton files):

https://github.com/lexnederbragt/den...leave_pairs.py

nposnien · 03-06-2012, 05:13 AM

First of all, thanks for the answers!

@ LizBent: Can I use the script for data that has been processed using CASAVA 1.8? In the discussion you added a link to, it is proposed to replace

f_suffix = "/1"
r_suffix = "/2"

with

f_suffix = ""
r_suffix = ""

My question is: How are the pairs identified then?

LizBent · 03-06-2012, 05:25 AM

No idea, you might want to ask the original script writer, who is cited in the comments at the top of the script (and there is also a reference to another SeqAnswers thread there that might answer your question). Sorry I can't help.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Splitting NuGen barcodes from paired-end sequences	ester	Bioinformatics	10	10-03-2011 10:36 AM
RNA-seq: read counts in single- vs paired-end sequences	fbarreto	Illumina/Solexa	4	08-03-2011 06:19 PM
SAMTool tview problem for paired-end sequences mapped by BFAST	NanYu	SOLiD	5	06-07-2011 02:08 PM
Velvet insert length on Illumina NGS Paired end reads	sari_khaleel	Illumina/Solexa	0	10-29-2010 09:12 AM
the oligonucleotide sequences of P7 and P5 on paired-end flow cells	lhemivw	Illumina/Solexa	0	08-20-2010 02:52 AM

01-27-2012, 06:23 AM	#1
LizBent Member Location: Guelph, Ontario, Canada Join Date: Jan 2012 Posts: 31	Velvet paired end after some sequences removed? Hi, After trimming and filtering my Illumina sequences (paired end 100 bp) for quality, I am left with files that probably don't have the same number of sequences in them any longer. This will be a problem if I try to interleave the files and use them as input for Velvet, right? Would it be better to: 1. Don't remove any sequences from my files, even if very short or 0 bases, so they can be interleaved 2. Remove sequences and use the (smaller) files in Velvet but not as paired-end reads Are there other options I'm not aware of? Thanks

01-27-2012, 10:22 AM	#2
jjohnson Member Location: Washington DC Metro Area Join Date: Aug 2009 Posts: 20	I would bin the valid pairs and singletons (those with mates removed due to quality trimming/filtering) into 2 separate fastq files. Velvet can accept mutiple files and then you can paramertize around the files (such as specifying insert sizes for mates file, etc). i.e. velveth Assem 35 -shortPaired -fasta pe_lib1.fasta -short3 se_lib1.fa __________________ Justin H. Johnson \| Twitter: @BioInfo \| LinkedIn: http://bit.ly/LIJHJ \| EdgeBio

03-05-2012, 01:42 PM	#4
rahularjun86 Member Location: Frankfurt(M), Germany Join Date: Jan 2011 Posts: 58	Dear nposnien, you can use Sickle tool (https://github.com/najoshi/sickle). You only need to input the pair fastq files, and other parameters (scoring system used, quality score to keep and length cutoff etc.), and it will generate the paired and singleton files. If you want to filter out reads with N's, Just replace the whole sequence with N and quality with #, then set Sickle length and quality values. This way it will filter out reads with N's. Best wishes, Rahul __________________ Rahul Sharma, Ph.D Frankfurt am Main, Germany

03-05-2012, 01:29 PM	#3
nposnien Member Location: Göttingen, Germany Join Date: May 2011 Posts: 13	Hi, we are facing the same problem at the moment. We will have uneven files (one for each pair) after trimming/filtering. My question is, if there is a script/program out there that would find the mates in in two different files (or in one file if I would merge/shuffle the files prior to trimming/filtering) and bins the unpaired reads into an extra file? Any help is highly appreciated!

03-06-2012, 12:21 AM	#5
LizBent Member Location: Guelph, Ontario, Canada Join Date: Jan 2012 Posts: 31	This script may be useful for interleaving pairs for Velvet (and generating non-paired singleton files): https://github.com/lexnederbragt/den...leave_pairs.py

03-06-2012, 05:13 AM	#6
nposnien Member Location: Göttingen, Germany Join Date: May 2011 Posts: 13	First of all, thanks for the answers! @ LizBent: Can I use the script for data that has been processed using CASAVA 1.8? In the discussion you added a link to, it is proposed to replace f_suffix = "/1" r_suffix = "/2" with f_suffix = "" r_suffix = "" My question is: How are the pairs identified then?

03-06-2012, 05:25 AM	#7
LizBent Member Location: Guelph, Ontario, Canada Join Date: Jan 2012 Posts: 31	No idea, you might want to ask the original script writer, who is cited in the comments at the top of the script (and there is also a reference to another SeqAnswers thread there that might answer your question). Sorry I can't help.