You have to find SNP's and then do a very high fidelity RT PCR and sequence the products. Once you have done that you should be able to tell the different lines from the sequence traces.

There are several ways but all require that you previously identified the SNP in your coding sequence taht will allow you to distinguish the alleles. Here are the techniques that I experienced:

- RNASeq and map the reads on polymorphic sites. Cutting edge technique, Expensive but exhaustive and requires high bioinformatic skills

- Pyrosequencing using the Qiagen Pyromark Q24 for instance. Expensive and limited in the number of SNPs to analyze. A standard in animal field

- qPCR. Technically difficult. It only worked in our hands when we designed each primer on a polymorphic sites. that means that you need 2 SNP very close. In our case that was not a problem.

Hope it helps,

Nicolas

Thank you all for your replies. I have a set of previously identified SNPs and RNA Seq data for the same individuals under two different sets of conditions. In literature, I found too many approaches to get to the allele specific expression which leads to confusion. So I would like to know how should I proceed from here? A very general work flow would be helpful to decide which approach I should take.

This is a basic outline of a workflow I have been using to generate ASE data:

1) Phase genotype data and impute (especially the genic regions) to get haplotype genomes for each subject. Imputation will help greatly if you retain high confidence imputed SNPs, and it more than doubled the number of allele-specific informative reads I had per subject. You could also include RNA-Seq called variants if you wanted.

2) Get the uniquely mappable haplotype genome reads for each haplotype. There is a function in the R package asSeq which will generate these BAMs for you, although it will require some additional processing of the phased genotype data.

3) Take your haplotype specific BAMs and obtain gene counts using your favorite feature counting software (I used Htseq).

Any downstream analysis will be dependent on your study design, etc.

Hope this helps.