Rna-Seq Mrnas Expressed In Only One Sample Set

Question

[Sudeep]

Hello all, In our rna-seq analysis we found a few mRNAs that are expressed only in one sample set (ie only in non-control sample set), does it make sense to go behind these mRNA (given that their appearance in only one sample set is not a sequencing anomaly) ? I was looking (googling) to see if anybody has done that and got some interesting results, but I couldn't find anything. Does anybody have some suggestions or good articles to point to ?

Thank you in advance.

[Nicolas Rosewick]

how many reads for these mRNAs ?

Answer 1

[Damian Kao]

It's very biologically possible that a gene is expressed under one condition and not expressed/lowly expressed under another. You can always design primers and do a quick semi-quantitative PCR (just look at intensity of the gel bands) or go for a full blown qPCR. But before you go down that route, you should dig deeper into your data:

-Are you poly-A enriching/ribo depleting?

-Are you sure the transcript is a coding mRNA? Are you mapping against a set of known annotations?

-What's the coverage of this transcript? How many reads mapped to this transcript?

-What does the coverage look like? Are most of the reads stacking in one region?

-Are there duplication events of this transcript in other regions of the genome?

[Sudeep]

Thank you for this wonderful suggestion

Answer 2

[Rm]

I agree with @DK: also see if you can run "cuffdiff" on the samples and see how large FPKM values for the sample where it mRNAs are expressed...IF FPKM vales are high enough then go for qPCR etc....

[Sudeep]

@RM thank you for suggesting cuffdiff I am a bit new to these next generation sequencing field