How to determine high-quality somatic mutations

Question

[vivekruhela]

Hi,

I want to determine significant or high quality somatic mutations from NGS data. I have already done annotation and functional significant determination of mutations. But I am not getting any way to find high quality somatic mutations because almost all the tools like Mutect etc require control data but whichever project I am working, they don't provide any control data. So is there any way to find high quality somatic mutations.

Thanks.

Edit 1 : I am not getting any way to find....(Sorry for mistake)

Answer 1

[Bastien Hervé]

"But I am getting any way to find high quality somatic mutations because almost all the tools like Mutect etc require control data"

You will not have quality somatic mutations without control. Variant Calling tools as Mutect rely on this control data to call variations

[vivekruhela]

Are there any other tools for this task, other than Mutect.

[Chirag Nepal]

varscan2, somaticSnipper, virmid, strelka and many more.

[vivekruhela]

As I have clearly mentioned, I don't have any control data. And all of the names mentioned by you need control data, so can you tell me any tool which can do this without control data.

[Bastien Hervé]

I just found this : https://www.nature.com/articles/s41598-017-14896-7

But I don't have feedback on it, so becareful

[vivekruhela]

Thanks. I'll check it out. As this tool works when there is only control data.

[vivekruhela]

Hey, sorry for very late response again. The link sent by you for nature paper, author has used controlled data but the difference is author has used control data of difference patient and tumor data of different and then used optimization technique for fitness function stabilization. And that control data is not open source data. So can you tell me how and where to get open source controlled data for multiple myeloma. Thanks again.

[sutturka]

As far as I know, the latest version of Mutect within GATK4 supports somatic mutation calling without paired normals. All you need is germline SNP data (may be from dbSNP) and you can create a Panel of Normal (PON) if you have access to public data (BAM files from normal individuals sequenced with same technology as your data, probably Illumina). If you are working with human data, there may be pre-made PON available. Hope this helps.