Estimate on-target coverage for exome data

[pravee1216]

Hello,

I was looking for a one-stop tool which takes BAM and target capture regions to provide a detailed summary with coverage estimates on-target/off-target regions for exome-seq, but I didn't see any such as of now. (There are tools that generate coverage files such as GATK's depth_of_coverage, BEDtools, but none of them provide *summary* of on/off-target bases/%)

Does anyone have any experience / know any such tools?

Raj

[Bukowski]

You could use Picard "CollectMultipleMetrics" and "HSMetrics" which tends to cover most things people want to know

[pravee1216]

Thanks for the quick reply. I have seen these tools, but "HsMetrics" require a bait file in addition to the target .bed. I have the target file, but I have no idea how to get the bait file and moreover its importance here, as target .bed can estimate the required matrics?

Do you have any idea how to generate/download the bait file for particular exon capture experiment or is this provided by Agilent/Nimblegen?

Best,

Raj

[Heisman]

Thanks for the quick reply. I have seen these tools, but "HsMetrics" require a bait file in addition to the target .bed. I have the target file, but I have no idea how to get the bait file and moreover its importance here, as target .bed can estimate the required matrics?

Do you have any idea how to generate/download the bait file for particular exon capture experiment or is this provided by Agilent/Nimblegen?

Best,

Raj

I use the same file for baits and targets.

[dnusol]

You could also try NGSrich

http://sourceforge.net/projects/ngsrich/

[Bukowski]

Thanks for the quick reply. I have seen these tools, but "HsMetrics" require a bait file in addition to the target .bed. I have the target file, but I have no idea how to get the bait file and moreover its importance here, as target .bed can estimate the required matrics?

Do you have any idea how to generate/download the bait file for particular exon capture experiment or is this provided by Agilent/Nimblegen?

Best,

Raj

As Heisman said you could just use the same file. I don't because I have separate baits and targets files (for SureSelect you can get these from e-Array), but in practice the on-bait and on-target figures don't differ that much that I would care about it..

[pravee1216]

Thanks for the reply. I would try both NGSrich as well as Picard metrics. I hope NGSrich can summarize with coverage plots (as explained here http://seqanswers.com/forums/archive...p/t-10014.html)

[pravee1216]

NGSrich is really good, but the number of reads (i.e., # reads) output is quite confusing as it gives a very low count, compared to the original total reads/on-target. For eg: # of reads say 0.2 million out of 150 million total reads, which is wrong. Have you ever noticed this before (in version 0.7.6)?

[dnusol]

No, I have not noticed that problem using v0.7.5

[annblack]

NGSRich target reads low

Hi - I noticed the same thing and opened a bug w/ the developer. He has fixed the read length issue and posted a new version of the tool.

Thanks!

[swbarnes2]

A pretty simple way...count how many reads align to the whole genome, use BEDtools to filter the whole .bam with a .bed file of your targets. Then count how many reads are in the filtered .bed file.