newbler assembly of mated reads (FASTA format)

dpuiu · 01-21-2010, 08:30 AM

Hello,

I'm trying to use Newbler de novo to assemble mated reads in FASTA/QUAL format.
Do you know if there is an implicit name convention for the paired read names?

That is not a problem if the input is an sff file. It seems to me that the assembler identifies the linker and appends _left & _right to the 5' & 3' ends.

Also, is there a way to explicitly specify a library's mean/stdev?

Thanks.

flxlex · 01-21-2010, 11:27 PM

First, if these are 454 reads I would try to get my hands on the sff file(s). I have yet to try, but my guess is the same read dataset assembles better with sff file input (anybody tried this already?).

If you are stuck with fasta/qual, then you should figure out which reads are pairs. Then, follow the newbler manual (page 156 for version 2.3) for adjusting the fasta/qual headers. In the header, you will have to indicate which reads are pairs (and from which library they are).

It is not possible to 'instruct' newbler to use a certain insertsize mean/stdev. Newbler will first build contigs, then map all pairs to them and use those where the forward and reverse read map to the same contig for calculating the mean insert size.

Good luck,

flxlex

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
BWA using reads in FASTA format	aggp11	RNA Sequencing	7	08-11-2012 05:55 PM
How to Improve Newbler Assembly	shuang	Bioinformatics	2	09-13-2011 09:45 PM
help needed to retrieve fasta reads from fasta db	prashanthpnu	Bioinformatics	1	06-21-2011 05:59 AM
PubMed: Detecting copy number variation with mated short reads.	Newsbot!	Literature Watch	0	02-18-2011 11:00 AM
Newbler assembly - excluding specific reads?	JonB	454 Pyrosequencing	1	01-20-2010 01:35 PM

01-21-2010, 08:30 AM	#1
dpuiu Junior Member Location: MD Join Date: Dec 2008 Posts: 1	newbler assembly of mated reads (FASTA format) Hello, I'm trying to use Newbler de novo to assemble mated reads in FASTA/QUAL format. Do you know if there is an implicit name convention for the paired read names? That is not a problem if the input is an sff file. It seems to me that the assembler identifies the linker and appends _left & _right to the 5' & 3' ends. Also, is there a way to explicitly specify a library's mean/stdev? Thanks.

01-21-2010, 11:27 PM	#2
flxlex Moderator Location: Oslo, Norway Join Date: Nov 2008 Posts: 415	First, if these are 454 reads I would try to get my hands on the sff file(s). I have yet to try, but my guess is the same read dataset assembles better with sff file input (anybody tried this already?). If you are stuck with fasta/qual, then you should figure out which reads are pairs. Then, follow the newbler manual (page 156 for version 2.3) for adjusting the fasta/qual headers. In the header, you will have to indicate which reads are pairs (and from which library they are). It is not possible to 'instruct' newbler to use a certain insertsize mean/stdev. Newbler will first build contigs, then map all pairs to them and use those where the forward and reverse read map to the same contig for calculating the mean insert size. Good luck, flxlex