I have RNAseq reads (single-end, 49-bp), I would like to do differential expression analysis. I know the standard pipeline is some rnaseq aligners ( e.g. Tophat, soap etc,) followed by featureCounts/Cufflinks(with assembly)/HTSeq and EdgeR/DESeq.
My question is, 1) will it be appropriate to use BWA aligner instead of other standard RNAseq specific aligners? And 2) why there is lot of difference in percentage of mapping reads when you do alignment with BWA and other RNAseq specific aligners?