Hi all,
I'm new to RNAseq analysis and am facing what seems to be an unusual situation, so I would love any and all advice you folks may have. We ran a stranded illumina truseq experiment which was meant to be paired-end. However, after R1 was successfully generated we experienced technical difficulties w/the sequencer and R2 is mainly of no use.
Obviously the ideal move here would be to rerun the experiment, however this is not an option. So, the plan is to move forward treating R1 as a stranded single-end experiment - which brings me to my pair of questions: 1.) Is this approach methodologically sound or is it invalid to try to interpret R1 without R2 (barring the obvious caveat that we will suffer all the disadvantages single-end data have relative to paired-end data)
and 2.) What are the appropriate options to feed featureCounts with respect to strandedness and paired-endedness?
So far I've tried strandSpecific=0,strandSpecific=1 and strandSpecific=2 - my expectation given the protocol was that strandSpecific=2 (reverse stranded) would be appropriate, however the unstranded option (strandSpecific=0) gives double the assigned reads as either strandSpecific=1 or strandSpecific=2. Is this expected behavior?
I've also been leaving "isPairedEnd" as its default (isPairedEnd=FALSE), but the function seems to recognize that the data were meant to be paired-end and prints "Paired-end reads are included" as part of its output. Will this affect my results?
Thanks so much!