IGV not showing the variants detected by GATK and Varscan

Question

[a.james]

Hello All,

I have a variant detected from both tools VARSCAN and GATK but when I look for it using IGV I couldn't find it.

Example: I have a variant at ch1:20006 T>G in a gene supported with a frequency of 10 % from both GATK and VARSCAN, but when I view it on the same sample in IGV I can not find the the reads supporting the variant G. Would be great if someone could share there thoughts on this.. Or an alternate way to look into this? This happened with a few genes..

Answer 1

[EagleEye]

1. First make sure that the genome version loaded on IGV and the genome you used to find variants are same.

   If the genome versions are fine play with alignment quality.

2. If you want to look all reads on browser, make mapping quality threshold to zero.

   Go to preference -> Alignments -> mapping quality threshold

   

   If you still could not see the reads.

3. Make sure that the alignment file loaded on IGV is not post filtered with mapping quality after you used for variant calling (try using the original file get after alignment, without any post filtering).

I hope some of these solutions work.

Answer 2

[Lemire]

Try playing with the viewing options under the "Alignment" tab under Preferences (or similar for your OS). In particular, uncheck the "Downsample reads" box.

Answer 3

[John]

GATK does local realignment of some reads. Check out this for answers:

http://gatkforums.broadinstitute.org/discussion/6235/haplotype-caller-calls-variants-where-there-are-no-reads#latest

https://www.broadinstitute.org/gatk/guide/article?id=5484

Answer 4

[harold.smith.tarheel]

You can use SAMtools 'tview' of the sorted BAM file to visualize the reads at your chromosome/position of interest.

[a.james]

Both fromIGV and BAm file I could see the reads of interest but not the variant in IGV alone..

[EagleEye]

Can you rephrase the sentence ? You mean to say that you can see the reads from the sample on IGV, but the variants (reported from the tools) are missing on that location. Is that it?

If that is the case, yes GATK does realignment and you should look for the realigned reads

http://gatkforums.broadinstitute.org/discussion/1235/i-expect-to-see-a-variant-at-a-specific-site-but-its-not-getting-called

https://www.broadinstitute.org/gatk/guide/article?id=5484

I also had the same problem long back but I did not want the realignment and needed simple haplotype caller to capture indels. Since I did not have time to look into the solutions for GATK at that point of time, I used simple haplotype caller callled Freebayes (Not alignment-based variant caller). It worked perfectly well for my PARCLIP (kind of) data and captured T -> C transitions very well (Figure 1C in the below article). I have also tested this tool for different datasets (not yet published) and it works well for SNPs.

Just for your reference:

This pipeline as been implicated in our recent article in Nature communications:

MEG3 long noncoding RNA regulates the TGF-β pathway genes through formation of RNA-DNA triplex structures

http://www.nature.com/ncomms/2015/150724/ncomms8743/full/ncomms8743.html

Pipeline used in the article along with freebayes: https://github.com/santhilalsubhash/TransExtract_betaV1.2

Since you have tried with two different tools, you can also try this tool and it is simple to use. https://github.com/ekg/freebayes

[a.james]

Thank you and yes I can use it..My requirement is only to get variants from a targeted regions of genome and so I did it with two tools inorder to have better sensitivity..

Can you rephrase the sentence ? You mean to say that you can see the reads from the sample on IGV, but the variants (reported from the tools) are missing on that location. Is that it?

--its Yes and No; As I could get a variant at a specific position from both variant callers (GATK and VARSCAN), I just wanted to view it in IGV for the same sample on the same locus for that captured varinat.. Where its showing the reads supporting that position but no, variant as reported by tools..But as you mentioned base reclaibration is a step I also followed from the GATK pipeline.. But the Recalibration is not applicable for VARSCAN ..

[EagleEye]

Ok then try this tool and let us know if you still could not solve it.