SNP zygosity from RNA-Seq data (BAM file)

Question

[ChIP]

Hi,

I have a list of SNPs, and I know number of tags covering these SNPs.

What I intend to find further, is whether the SNPs are homozygous or heterozygous from RNA-Seq tag counts.

I have also computed the RNA-Seq tag counts for each of the SNP, like below:

bedtools multicov -bams RNA1.sort.bam RNA2.bam -bed Cordinate.SNP.txt

RNA-seq I have is a paired end.

Could anyone help, in figuring out if these SNPs are homo or heterozygous based upon the RNA-seq tags covering them.

Thank you

[Brian Bushnell]

You may be able to confidently call a het from RNA-seq data, but I doubt you can do that for homozygous because one copy might be shut off and not expressed.

Answer 1

[H.Hasani]

Well, some mutations are of epigenetic nature(e.g. mono-expressed genes). Therefore without a DNA, there is no way of telling whether the original locus is homo or hetero. In the same context, imprinted genes & LOH are another example for what seems on RNA as a mono (please note that these are all different cases).

Moreover, hetero mutations, are not necessarily coming from hetero locus! It could be that the sample is contaminated, or it is an indication for multi-colony sample.

Answer 2

[Shicheng Guo]

Suppose your reference genome is human, then call SNPs with your RNA-seq data and then compare the called SNPs with your Cordinate.SNP.txt. The accuracy should be evaluated before make further conclusions since imprinted genes and RNA-editing. Later, I will find a data test it and then I can share the accuracy performance