Is mixing samples (tissues from several mice in one replicate) OK for RNA-seq? and does it reduce the requirement for replicates?

Question

[NGS_enthusiast]

Hi, I was wondering if mixing samples (we probably have to, in order to have enough RNA) was OK? I was also wondering if it could increase reliability from few replicates, since lots of samples mixed in one replicate would help reduce variability? For example maybe 2 replicates of 5 mice each would be better than 4 replicates, each from 1 mouse. thanks a lot for your input, best regards

[ATpoint]

How much RNA are you getting out per animal? It is simple, if you don't get enough RNA for a per-mouse prep then yes mix it, since there is no alternative. It is not uncommon to do that, specially for assays like ChIP-seq where you need many cells or if material is super rare. There are today low-input RNA-seq kits though. You can use them with a few hundred or thousand cells, required quite some PCR amplification though.

[NGS_enthusiast]

yes thanks well I was pondering whether to try low input or keep high input and mix samples.

[ATpoint]

And what did you decide for eventually?

[NGS_enthusiast]

For this experiment, I decided to mix some samples as low input libraries are not yet available at the facility where we do our sequencing. But I'll try to change in the future.

Answer 1

[swbarnes2]

For example maybe 2 replicates of 5 mice each would be better than 4 replicates, each from 1 mouse.

No. Most of the point of replicates is to understand how much the gene naturally varies among the replicates; averaging them all together destroys that. 4 separate replicates is absolutely better than smooshing 2 groups of 5 mice into two combined samples.