RNASeq Analysis: data from different library

Question

[Matteo]

Hello everyone

I’m trying to figure out if it’s possible to analyze NGS data from datasets obtained using different library prep strategies. Specifically, I want to know if there’s a way to combine a total RNA-seq dataset with one using poly-A selection for determining DEGs.

Thanks
Matteo

[i.sudbery]

If you are getting one condition from the total RNA-seq, and the other condition from the Poly-A (e.g. the control from total, and the treatment from poly-A), then the answer is no.

[Matteo]

Thanks for your answer. No no, I'm going to have the same conditions over the two groups:

• total RNASeq: 50% cond A 50% cond B
• Poly-A: 50% cond A 50% cond B

Maybe using only the transcriptome reference for the transcript quantification and, of course, adjusting for this batch effect in the design formula or removing it with sva or similar

[swbarnes2]

Since you've balanced library prep method with experimental conditions, it will work. But I'd examine the PCA of all the samples to see how much of an impact it has, and be sure to add 'batch' to your design when looking for DE genes.

[i.sudbery]

This. No need to look for hidden effects or use NOI-seq - the effects are not hidden, you know exactly where they are!

[Arby]

You can also use something like NOISeq to correct hidden effects