detecting allele specific expression of a gene at a specific locus without having any information about parental genomes

Question

[prasundutta87]

Hi,

I have RNAseq data of a water buffalo based on which I have got SNV and INDEL information in a VCF file (all variants annotated as well). Is it possible to decipher allele specific expression information from such data? Unfortunately, I do not have their parental genome information. Any logical/methodological answers are welcome. I am doing such analysis for the first time.

Some more details: I have got read counts for my reference allele and alternate allele which I got after using allelecounter.py program (github). I am stuck after this step. There are a variety of papers which infers to a lot of methods but I am not sure which method to go ahead with. I know that I can infer about allelic imbalance by doing a binomial test (used python scipy package for that). But, if I do not have any parental genomes, is it possible to determine from which chromosome has a particular allele has come?

Answer 1

[kirannbishwa01]

Check this software: https://www.nature.com/articles/srep21134 i.e GeneiASE, but I have not used it personally. The big deal in ASE analyses is that we are making assumptions that both the alleles have equal likelyhood of mapping, but the actual scenario is different (there is a mapping bias). So, some methods to actually account for this is to: 1) create a phased diploid genome 2) create a haploid genome to allow ambiguity matching 3) or you need to find a custom way for handling your mapping bias.

After that it's just a applying proper statistics to the calculated (observed) count data. There are lots of different tools, software to test ASE.

Thanks,

[prasundutta87]

Hi..thanks for your answer..I have come across this tool after I used and got results using MBASED, another tool that gives gene based ASE without requiring parental genomes or phasing information..

[kirannbishwa01]

yeah, I have come across MBASED, but I think its for comparing ASE between two replicates of same sample. It may be otherwise, but I would look into the assumptions made by the tool/author in detail, before relying on ASE report.

[prasundutta87]

I have performed the one sample analysis and have followed their protocol from their manual..I have been having discussions with the author (Oleg) and have shared my methodology with him and all seems to be on track..

[kirannbishwa01]

That sounds good that it is working. :)