questions: de-novo transcriptome assembly

The Megakaryocyte_rep2 data is much smaller than the other samples. Is this intended? In https://www.ncbi.nlm.nih.gov/sra/?term=SRR549358 the original data of the two samples is about the same size. It seems unlikely that down sampling created this bias. Maybe some error happened while downloading / uploading?

StringTie is can estimate transcript abundances on its own. Then what is the intension of using featurecounts+DeSeq2 instead of directly using DeSeq2? If Stringtie is given the gff then one can configure to output DeSeq2 compatible output. I guess its just that when the tutorial was created this possibility was not yet available...