Go Back   SEQanswers > Bioinformatics > Bioinformatics

Similar Threads
Thread Thread Starter Forum Replies Last Post
Merging genomic Assemblies rahularjun86 De novo discovery 0 02-03-2012 02:08 AM
Align genomic DNA sequence to protein database? rdu Bioinformatics 0 11-03-2011 07:31 PM
Can BWA align reads to references that are even shorter? asiangg Bioinformatics 4 04-06-2011 11:06 AM
Align a whole genome with BWA lyz1030 Bioinformatics 0 03-17-2011 12:50 PM
Identification of Genomic Repeats and Sample Contamination in Assemblies of 454 Pyros flxlex Literature Watch 0 01-14-2010 01:31 AM

Thread Tools
Old 01-27-2010, 06:10 PM   #1
Location: St. Louis

Join Date: Dec 2009
Posts: 74
Default Using bwa to align cDNAs to genomic assemblies?

I've just started playing with the long read algorithm for bwa (dbwtsw), and I was wondering if it would be useful for aligning 454 or Sanger length cDNA reads back to a genomic assembly.

Back in the old days, I'd use wu-blast for this kind of thing, its ability to sort HSPs into alignment groups, and report only the most likely grouping of HSPs using topcomboN was tremendously helpful to me. I could layer my cDNAs onto genomic sequence, and in most cases blast could piece together alignments broken even by large introns.

I was wondering how bwa behaved in cases where a long (454-length) cDNA read spanned an intron in a genomic reference. Will this just drop the score so much that bwa won't even realize its an alignment? Or will it show the partial alignments of each end of the cDNA read to the seperated positions on the genomic reference in true 'local' fashion?

If bwa is not able to do this, can anyone suggest a modern aligner that can? Or should I just stick with blast and/or blat?
jmartin is offline   Reply With Quote
Old 01-27-2010, 06:55 PM   #2
Junior Member
Location: Perth, Western Australia

Join Date: Jan 2009
Posts: 6

I've only played around with bwa. While excellent for mapping reads with a few indels, there are serious performance hits for allowing more than a handful. I'm pretty sure that these are fundamental limitations that stem from the way that bwa constructs the reference hash. These limitations are the trade-off you accept to get the speed of the burrows-wheeler approaches.
I can't suggest a really great tool for the job, perhaps someone else can chime in?
robsyme is offline   Reply With Quote
Old 01-27-2010, 07:06 PM   #3
Nils Homer
nilshomer's Avatar
Location: Boston, MA, USA

Join Date: Nov 2008
Posts: 1,285

You may be confusing the BWA "short read" alignment algorithm with "long read" alignment tool algorithm, which has just been published. The "long read" algorithm is meant for long queries (200bp-1Mb) and is comparable to BLAST/SSAHA2 algorithms. The sensitivity of the algorithm, based on the paper, is quite good with a nice speed-up over BLAT/SSAHA2.

I am not sure how many "introns" are allowed by "bwa dbwtsw" since in the paper they show sensitivity of chimeric reads that have one breakpoint. You can PM lh3 (the author) or you can simulate some cDNA reads and find out; post the results if you do simulate. Your best bet until it is determined if BWA can work and if you are not worried about speed is to use BLAST and then write a script to parse the results.
nilshomer is offline   Reply With Quote
Old 01-27-2010, 07:16 PM   #4
Junior Member
Location: Perth, Western Australia

Join Date: Jan 2009
Posts: 6

@nilshomer - My mistake, you're absolutely right, I misread the initial post.
robsyme is offline   Reply With Quote

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

All times are GMT -8. The time now is 10:13 PM.

Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2020, vBulletin Solutions, Inc.
Single Sign On provided by vBSSO