Divergent species and low mapping rate

Question

[snow4964]

I am trying to do a differential expression analysis to identify potential targets of interest in Bdelloidae species. I am working with RNA-Seq data of Philodina sp. however there is not much available sequence information for this genus. However Adineta (a different order) has a published information on ensembl. My plan, at this moment, is to compare the expression levels of my data to that of published Adineta.

I have mapped my data (Philodina) to the Adineta reference available on ensembl using salmon. However the mapping rate of my samples to the Adineta reference were consistently >10%. Additionally, I also mapped similarly divergent published RNA Bdelloidae data (from Rotaria, a third order) and had similar mapping rates to that of my samples.

Is this to be expect with the mapping of diverged species? Although the mapping rates of my samples are very low, can my salmon output still be used?

Answer 1

[cfos4698]

Ideally you'd want much higher mapping rates than that. There is likely a lot of information in your RNASeq reads that is being discarded due to not mapping to what seems to be a relatively distant reference genome/transcriptome. Have you tried doing a de novo assembly, annotating the assembly, then mapping against your de novo reference for differential expression? That approach would likely result in a much higher mapping rate and give you more useful results.

[snow4964]

Thanks for the quick response!

I did try mapping my samples (Philodina) and the Adineta samples to a de novo assembled, with Trinity, version of my data. However, if I remember right, the mapping rate was then low for the Adineta samples opposed to my samples. I will try it again to see if I may have misread or overlooked the mapping rate, but if y memory serves me right it was still below 10%.

It may be important to note that I have been using the Adineta CDS from Ensembl. Though at this point I am considering using the Adineta cDNA assembly to see if that increases my mapping rates at all...

Update: I mapped the available Adineta reads to my assembled transcriptome. The mapping rate was slightly higher, but still below 10%