Too many intronic variants in RNASeq, is this ok?

Question

[Sharon]

Hi every one

I did GATK somatic mutation pipeline, then I annotated the resulted mutect2 vcf with VEP. 94.32% of mutect2.vcf variants are annotated as intergenic and intronic, and the rest is exonic or ncRNA exonic, is this normal with rnaseq variant calling and somatics? I have many reads discarded by mutect2 though.

Thanks

[karl.stamm]

Look more closely at your data. Depending on the alignment methods, maybe you have bits of RNA-transcript hanging off the end of an exon, and your tools see intronic mutations. I have seen piles of artifact intronic variants when the read was not trimmed from the end of an exon. It's not really there, but a problem with alignment.

[Sharon]

Thanks karl so much. I will check the alignments with IGV.

[WouterDeCoster]

Are you following guidelines for this? I don't think gatk somatic is appropriate for RNA-seq but might be wrong...

[Sharon]

I am using GATK and mutect2, why it is not appropriate?

[WouterDeCoster]

I believe those are optimised for DNA sequencing such as exome sequencing.

[pereyra.s]

Sharon, what did you conclude from your data? I am having a similar problem with my rnaseq variant calling analysis (aprox 70% intergenic or intronic).