I am planning to perform RNAseq on mouse samples and I am trying to figure out which would be the best reads length to cover more than one exon-exon junction in order to get isoform. In the literature I found 75bp was used for human sample, would it be fine also for mouse? or I should go for 100bp?
Either should be fine. Longer reads give more precise mapping (in general).
You may want to consider Iso-Seq on the PacBio platform (http://www.pacb.com/applications/rna-sequencing/). I've heard generally positive reviews.
That will depend on his budget... but it's indeed pretty cool.
Yeah, and they won't get any useful (or at least of very limited use) quantification data. Iso-Seq may be better for transcript discovery in a largely unexplored transcriptomes rather than something like mouse.
I would indeed guess that it's not as quantitative as established illumina RNA-seq. But I disagree that it wouldn't be an added value to characterize the transcriptome of mouse. I'm sure there are many transcripts we haven't discovered or properly annotated yet, definitely in lesser studied/accessible tissues. But we don't know what OP exactly wants to get out of this analysis :p
My mouse RNA-seq was 90 bp. It depends of your resources and what you want.
Potential useful resources: http://core-genomics.blogspot.be/2016/07/rna-seq-advice-from-illumina.html and http://mkt.illumina.com/GM-CPBU-RNASEQ-PPC-Q12016_LP-StaticwithForm.html?scid=2016213PPC13