Why is stranded RNA seq superior to unstranded? From this seqanswers thread I understand that the stranded sample prep protocol makes sure that only the sense strand is sequenced (the information you need?) whereas with unstranded the complementary is sequenced as well (you dont need that?) So does that mean that when using the unstranded protocol about 50% of the data is useless and thats Why stranded is better?
The reason stranded is better than unstranded is that you can tell which strand your RNA is being transcribed from (kind of obvious). There is a LOT of overlapping (antisense) transcription in complex mammalian genomes: for example, if you have a look at the BDNF locus in humans (UCSC hg19 genome browser coordinates chr11:27,671,365-27,684,616) you will see that there is (in the "middle") a region which has overlapping exons of the BDNF protein-coding gene on the (-) strand and BDNF-AS regulatory RNA on the (+) strand. With unstranded sequencing, where you have reads mapping to that area, you would have no idea which of the two transcripts was being upregulated - the mRNA or the ncRNA, and this would obviously affect what you think is happening at the biological level.
The way this had been "gotten around" in unstranded sequencing is by the implicit assumption that every transcript has some "unique" region that is characteristic only of it, which will be used for counting how many reads map to it and doing differential expression calculation.
Also it's useful to remember that the antisense transcriptome has not been characterized very well because stranded protocols haven't been around for very long or used quite widely.
But when using unstranded it's not that you "discard" 50% of your data - it's that the sequence you see in the read may correspond to either the 1st or 2nd cDNA strand and you have NO IDEA which one it is.
Hope this helps!