How to deal with no calls from GATK Unifiedgenotyper for indels

audqf · 01-30-2012, 02:47 PM

I am working with 454 single-end long-read data for a region on chr6, using hg19 and dbSNP build 135 as references. GATK Unifiedgenotyper with the option -glm INDEL returns no calls on realigned+recalibrated bam file and on recalibrated bam file. Using additional arguments, such as --output_mode EMIT_ALL_SITES, or setting call/emit thresholds, does not make a difference, either.

Earlier posts on this forum (http://seqanswers.com/forums/showthread.php?t=15634) indicate that a high sequencing error rate may be the cause.

So I wonder:
1. If resetting the sequencing error rate is the solution, how do I do it? Is this error rate reported in the bam file?

2. Are there other ways of tweaking GATK to deal with this problem?

Thanks for your help!

ulz_peter · 01-30-2012, 11:43 PM

I once tried to twaek GATK to output Indels in 454 data, but I think they turned that utility off as soon as you specify the PL:454 (or maybe it was PL:ROCHE, I don't remember) tag. In case you find a solution, I'd be happy to it

audqf · 02-01-2012, 03:53 PM

In case ulz_peter and others are interested, it turns out that GATK UG does not make indel calls for 454 data.

I also posted the same question here:
http://getsatisfaction.com/gsa/topic...#reply_7912928

(btw, this is a pretty good place to have GATK-related questions answered)

Now the wiki page for GATK UG clearly states this limitation:
http://www.broadinstitute.org/gsa/wi...fied_genotyper

Hope this helps. I'm trying samtools for indel calling now.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
GATK UnifiedGenotyper not calling any variants	krobasky	Bioinformatics	7	10-25-2012 11:19 PM
Trouble with making indel calls using UnifiedGenotyper	oiiio	Bioinformatics	0	11-04-2011 01:55 PM
GATK update changes UnifiedGenotyper output	doc.ramses	Bioinformatics	2	10-18-2011 07:41 AM
GATK UnifiedGenotyper calling way too many SNPs in vcf	swbarnes2	Bioinformatics	0	08-17-2011 02:33 PM
missing header information in bam cause GATK unifiedgenotyper fail	foxyg	Bioinformatics	3	11-11-2010 10:37 AM

01-30-2012, 02:47 PM	#1
audqf Junior Member Location: Chicago Join Date: Jan 2012 Posts: 2	How to deal with no calls from GATK Unifiedgenotyper for indels I am working with 454 single-end long-read data for a region on chr6, using hg19 and dbSNP build 135 as references. GATK Unifiedgenotyper with the option -glm INDEL returns no calls on realigned+recalibrated bam file and on recalibrated bam file. Using additional arguments, such as --output_mode EMIT_ALL_SITES, or setting call/emit thresholds, does not make a difference, either. Earlier posts on this forum (http://seqanswers.com/forums/showthread.php?t=15634) indicate that a high sequencing error rate may be the cause. So I wonder: 1. If resetting the sequencing error rate is the solution, how do I do it? Is this error rate reported in the bam file? 2. Are there other ways of tweaking GATK to deal with this problem? Thanks for your help!

01-30-2012, 11:43 PM	#2
ulz_peter Senior Member Location: Graz, Austria Join Date: Feb 2010 Posts: 219	I once tried to twaek GATK to output Indels in 454 data, but I think they turned that utility off as soon as you specify the PL:454 (or maybe it was PL:ROCHE, I don't remember) tag. In case you find a solution, I'd be happy to it

02-01-2012, 03:53 PM	#3
audqf Junior Member Location: Chicago Join Date: Jan 2012 Posts: 2	In case ulz_peter and others are interested, it turns out that GATK UG does not make indel calls for 454 data. I also posted the same question here: http://getsatisfaction.com/gsa/topic...#reply_7912928 (btw, this is a pretty good place to have GATK-related questions answered) Now the wiki page for GATK UG clearly states this limitation: http://www.broadinstitute.org/gsa/wi...fied_genotyper Hope this helps. I'm trying samtools for indel calling now.