Maximum variation in biological replicates

Question

[Prasad]

Hello,

What is the maximum difference allowed in biological replicates. For ex. I have three biological replicates called A1, A2, and A3, suppose i have done rnaseq and i could identify 10K in A1, 5K in A2 and 15K expressed transcripts. Is this difference allowed in biological replicates? Are there any articles to check max variations in biological or technical replicates

any suggestions are appreciated. Thanks

Answer 1

[Devon Ryan]

The whole point of replicates is to allow you to judge biological variation. There's no maximum value.

If you're wondering if you have an outlier sample, then the best method is to make a PCA plot or do some hierarchical clustering and then show that to a toddler. If you ask the toddler, "which of these points/samples looks out of place?" and the toddler has to think about it then your sample isn't enough of an outlier to exclude*.

*A similar method can be used to find TADs in HiC data or ChIPseq peaks, but then you need more toddlers.

[Prasad]

thanks for the reply.

then the best method is to make a PCA plot or do some hierarchical clustering

is the best method to show how replicates are behaving. In my case i have 1 replicate for each condition i.e. A1 A2 for control and B1 B2 for treated. I will do the clustering. any article/s which tells variation in terms of % would be really helpful