feature count strand explained

Question

[popayekid55]

Hello, I am very new to rna-seq. I have got the RNA-seq data for bacteria (Hiseq 150*2). From lab it was told that its a directional library (NEB ultra directional kit). Now i assume the data is from Sense/Forward strand.

I have used featureCount with -s 1 and only 1% of the reads were assigned to feature. Then i tried with -s 2 (reverse strand) and >96% of reads were assigned. Could someone explain strand concept in feature count.

Thanks

Answer 1

[lshepard]

Hi, check out the following links for the description of strand specificity. From the output from featurecounts, looks like your first strand was the reverse (ISR in the first link). They really helped me:

http://salmon.readthedocs.io/en/latest/library_type.html

http://onetipperday.sterding.com/2012/07/how-to-tell-which-library-type-to-use.html

Answer 2

[WouterDeCoster]

You can look at the strandedness of your data in a browser such as igv. Are the reads in the same direction as your genes or in the reverse direction? For most recent stranded kits the reads are most often reverse stranded, which as you observed gave the best result.

[popayekid55]

For most recent stranded kits the reads are most often reverse stranded,

Why is it so??

[WouterDeCoster]

Because that's the way the library prep works, the second strand is not amplified (in the case of dUTP method).

[popayekid55]

thanks. If i count the data with unstranded option how does that impact the?

[WouterDeCoster]

In all the excitement it seems you lost a word of your sentence.

[popayekid55]

If i count the data with unstranded option how does that impact the differential expression analysis/any down stream analysis?

[WouterDeCoster]

Exons which are antisense will get counted incorrectly or not at all because the reads cannot unambiguously get attributed to one of both transcripts.