RNA-seq experiment design

[Kulvait]

Hi,
we have mouse models with knock out of some genes. Basically we have at least 3 different mouse genotypes with additional problems with variability of potential biological multiplicates in a genotype group (male/female, relatives, age). We are considering RNAseq that we has not done before so my group have no experience in it. Any help in the design phase will be highly appreciated.

We would like to perform RNA-seq in order to answer these questions:

• How the knock out genes affect resulting rna transcripts forms?
• Are there also differences in rRNA or microRNA?
• We are partly interested in differential mRNA expression.

I would like to ask on this forum few questions regarding experimental design.

• Do you think that sequencing whole RNA and not only mRNA is reasonable? What to expect if we do that? Do you have experience with that and are there some analytical approaches for such data?
• We will have paired end rapid run sequencing on Illumina Hi-Seq 2000. What the sequencing depth should be? And if we want to achieve given sequencing depth (i.e. 1000x) how much sequencing capacity should we allocate per sample?
• Is it sufficient to have biological replicates or technical replicates are also needed? If we have three genotypes how many samples per genotype is sufficient to perform any meaningful analysis of differential transcript forms in different genotypes?

Thank you for answers,
Vojtech.

[Genetic Librarian]

Hi Voijtech,

first, I would recomment you to read this for some more background:

https://genohub.com/next-generation-sequencing-guide/

Second: Getting good data for miRNA, rRNA and mRNA in a single experiment is not reasonable.

If you sequence total RNA, 98% of your reads will be rRNAs.
In my opinion, you would need three library types (miRNA, total RNA without enrichment and mRNA) to check for all types.