This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Multi-Sample Indel Realignment Using Gatk

Is there a way to use the GATK Indel Realigner tool with a two-samples-input?

The issue is that in my normal sample, several reads are mapped containing SNPs at specific positions, and ONE read having an insertion and no SNPs anymore. As aspected, the GATK Indel Realigner removes the SNPs from the other reads and includes the corresponding insertion. So far so good, but my tumor sample does NOT contain any read with the insertion. Only wild-type reads on the one hand and reads having the mentioned SNPs on the other hand. So, the GATK Indel Realigner does no realignment on these reads. As a result I have somatic SNPs called, as they occur in the tumor sample but not in the normal sample. Without the Indel Realigner I would not get these false positives, as they would occur on both samples...

If I could use both samples coincidentally as the input, the algorithm could test the second sample as well for realignment if an appropriate position is found in one sample.

Does anyone know an answer to this problem? Will this eventually be included in upcoming versions of GATK?

Thanks a lot in advance,

Cheers

indel gatk
• 4,397 views • link
updated by Biostar Biostar • written by Sebastian • Sebastian

that's not clear to me. could you please add a (masked) IGV screenshot of your regions w/without indel-realigner ?

• 0 views • link
written by Pierre Lindenbaum Pierre Lindenbaum

1 answer

GATK does accept merged BAM files, as far as samples are well defined at the BAM header. also GATK accepts several BAM inputs using the "-I" option several times, as far as the different samples have different RG labels at their headers, so that should work for you.

although I haven't used this feature myself, I understand from the documentation that the information from other samples would be used to improve the algorithm for a particular one in the same way the variant detection results are improved when performing calls on several samples at the same time. still, the best practices document does not encourage users to use it in this way for regular basis, but I understand that your situation justifies such usage.

• 0 views • link
written by Jorge Amigo Jorge Amigo

Log in to answer this question.

More posts like this