computational SNP Identification/Analysis using RNASeq data

[afiroz]

Dear All,

I want to learn SNP Identification/Analysis using RNASeq data, I will be thankful for the computational protocol of the same.
Thanks

[atcghelix]

Do you have a reference genome available? If so the below might not be relevant to you.

There are some challenges to SNP discovery in RNA-seq data that are still being worked out, as far as I know, especially when you don't have a reference genome. Basically, when you assemble RNA-seq data you expect to get many different isoforms of the same gene. These isoforms are sometimes grouped with paralogues which can be difficult to distinguish from isoforms. When it comes time to calling SNPs, you want to find some way to reduce the redundancy of the de novo assembly to more accurately represent the underlying biology of a single genomic origin for multiple transcripts. If you simply cluster using something like CD-hit-est this will do things like collapse paralogues too.

This doesn't mean it's impossible, and I'd love to hear others' input. I'm trying to do this in a few species with no close reference genomes as well.

I've found this paper to be helpful: http://onlinelibrary.wiley.com/doi/1...998.12077/full

[afiroz]

Thanks for reply... yes If having reference genome..... what would be the protocol for the SNP identification and analysis... also which tools is most useful for same. thanks again for valuable response/reply.
firoz

[atcghelix]

Do you know where your exon junctions are for your genome?

[shi]

Thanks for reply... yes If having reference genome..... what would be the protocol for the SNP identification and analysis... also which tools is most useful for same. thanks again for valuable response/reply.
firoz

You may try the Subread package which includes an aligner (subjunc) that performs full alignments of RNA-seq reads and also a SNP caller (exactSNP).

[afiroz]

Dont know the exon junction

[shi]

You dont need to know the exon junctions. The aligner will detect them for you.

[afiroz]

Thanks for valuable reply. am trying subread packege (aligner) and will let u know further.. thanks again