Doubts about GATK "raw data processing" step for SOliD exome data

jorgebm · 06-14-2012, 07:56 AM

Hello,

I just received SOLiD 75bp fragment (SE) exome data. I haven't previosly worked with SOLiD data and I'm not sure the suitable way to handle them for a further GATK based variant detection. In concrete I have doubts about:

- filtering PCR duplicated

I'm undecided to filter or not them using picard markduplicates. Being SingleEnd data I'm inclined not to do it to avoid lose "true" reads (not duplicates) at expense of generate "dirty" variant calls. What's you opinion?.

- base recalibration

Due to special caracteristics of colorspace read seems to be unnecessary to make base recalibration. I checked "1000 Genomes" paper (http://www.nature.com/nature/journal...ture09534.html) where based recalibration was avoided for SOLiD data. By other hand, in GATK documentation (http://www.broadinstitute.org/gsa/wi..._recalibration) "PrimerRoundCovariate" covariate is described as "The primer round for this base (only meaningful for SOLiD reads).". I'm confused and do not know if the recalibration of bases is a recommended step for SOLiD.
Any advice?

I would appreciate your help.

Thank you

szilva · 06-18-2012, 05:00 AM

Actually I also have similar doubts (though working with paired data). Do you have any progress or hints on this topic?

jorgebm · 06-18-2012, 06:17 AM

Unfortunately, don't. But I think If you've PairEnd data you could safetly remove PCR duplicates. However, I neved did it for SOLiD.

About "base reclaibration" I'm going to try it. I'm waiting for finishing alignment step to try it. I'll let you know then.

By other hand, I'm using BFAST to align to colospace but it is extremely slow. What alignerare you using?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Bioinformatician "Data Finish & Quality Control"	amb	Academic/Non-Profit Jobs	0	06-08-2012 01:19 AM
raw data of paper "The evolution of gene expression levels in mammalian organs"	rzhang	Illumina/Solexa	5	02-27-2012 06:08 PM
Option "calmd"; Reporting indels and Somatic mutations for Whole Exome Seq data:	angerusso	Bioinformatics	0	01-10-2012 04:32 PM
human genome short read data for "Do it yourself genetic testing"	attilav	General	2	08-22-2011 01:59 PM
normalizing RNA-seq data to "unique transcript length" instead of "transcript length"	lmc	Bioinformatics	2	06-23-2010 11:45 AM

06-14-2012, 07:56 AM	#1
jorgebm Member Location: Spain Join Date: Feb 2010 Posts: 18	Doubts about GATK "raw data processing" step for SOliD exome data Hello, I just received SOLiD 75bp fragment (SE) exome data. I haven't previosly worked with SOLiD data and I'm not sure the suitable way to handle them for a further GATK based variant detection. In concrete I have doubts about: - filtering PCR duplicated I'm undecided to filter or not them using picard markduplicates. Being SingleEnd data I'm inclined not to do it to avoid lose "true" reads (not duplicates) at expense of generate "dirty" variant calls. What's you opinion?. - base recalibration Due to special caracteristics of colorspace read seems to be unnecessary to make base recalibration. I checked "1000 Genomes" paper (http://www.nature.com/nature/journal...ture09534.html) where based recalibration was avoided for SOLiD data. By other hand, in GATK documentation (http://www.broadinstitute.org/gsa/wi..._recalibration) "PrimerRoundCovariate" covariate is described as "The primer round for this base (only meaningful for SOLiD reads).". I'm confused and do not know if the recalibration of bases is a recommended step for SOLiD. Any advice? I would appreciate your help. Thank you

06-18-2012, 05:00 AM	#2
szilva Member Location: Hungary Join Date: Aug 2009 Posts: 16	Actually I also have similar doubts (though working with paired data). Do you have any progress or hints on this topic?

06-18-2012, 06:17 AM	#3
jorgebm Member Location: Spain Join Date: Feb 2010 Posts: 18	Unfortunately, don't. But I think If you've PairEnd data you could safetly remove PCR duplicates. However, I neved did it for SOLiD. About "base reclaibration" I'm going to try it. I'm waiting for finishing alignment step to try it. I'll let you know then. By other hand, I'm using BFAST to align to colospace but it is extremely slow. What alignerare you using?