RNA-Seq Strand Bias

Question

[skbrimer]

Good afternoon collective hive brain,

I bring forth another head scratcher for you. I'm working with Ion Torrent generated RNA-Seq data and I have noticed that the vast majority of the reads that map back to the reference are the reverse strain. Is this normal? I get full coverage of the target ssRNA virus we are working with however even with cDNA creation would I not see more of a 50/50 split between reads?

I would like to see if how all the reads shake out, not just the ones that map. Does anyone know how to split a file of reads based on their directionality. I'm assuming this is possible since you can check for directionality of mate pairs but I do not normally work with mate paired data so I'm not sure how to do that. I figure this should be doable either with the the fastq file or the sam file, but I can not find in the documentation which flag or character is the one that tells you the direction.

Thanks in advance,

Sean

[Carlo Yague]

I can't really answer your question, just be aware that you can't determine the strand of an unmapped read. You get the directionality only after the mapping.

Answer 1

[Michael]

That's normal for your target and a strand specific protocol, hint what does ssRNA virus stand for? I guess you have a -ssRNA virus?

[skbrimer]

Thanks Michael,

I love how the internet keeps me humble. I'm actually working with a +ssRNA virus and after your answer I went back and re-read the protocol, since these are barcoded libraries the cDNA amp step is only performed with the 3' primer not both the 5' and 3' (as with the non barcoded protocol). I thought because the samples are amp'ed that I should see more of a 60/40 ratio but since they are only amp'ed from the 3' end it makes sense that I only get the reverse strand.

Sean