We plan to do an RNA-Seq analysis of knockdown and control samples in order to analyse differences in alternative splicing patterns between the two groups. The analysis should be done using R package SGSeq. We would now like to know which parameters we have to choose for the RNA-Seq experiment that allow us to do the analysis with SGSeq lateron.
We know that we need strand-specific sequencing with long paired end reads (100bp-150bp). We are however unsure about the sequencing depth. Do you have suggestions for that?

Thank you for your help,
best regards,
Andrea

Hi Andrea, 

Actually strand-specific sequencing is not a requirement (the package was written a few years ago when most available data sets were not strand-specific). Since SGSeq relies on split reads, longer reads are helpful. There is no requirement for a particular sequencing depth. More reads are better, especially when analyzing genes with low expression, but this has to be weighed against higher cost. Standard recommendations for transcriptome studies should be a good starting point. 

Best, 

Leonard

Dear Leonhard,

thanks a lot for your reply. Good to know that we don't necessarily require strand-specific sequencing. Concerning the sequencing depth, do you think that 50M reads for each sample with 3 biological replicates per group would be ok?

Best,

Andrea 

Hi Andrea, 

If you have the option to do strand-specific sequencing, this is still a good idea for transcriptome analysis (just not required for SGSeq). Read depth and replicates sound reasonable. 

Best, 

Leonard

Hi Leonard,

ok, I think we will then perform the experiment like that. Thanks a lot for your help.

Best,

Andrea