Hi All,

I am writing to clarify a technical aspect for differential expression analysis. For some practical reason, we could not have biological replicates. However, in order to perform mRNA-seq, we used technical replicates. Technical replicates here mean that: From 20 plants, leaf tissues were pooled together and powder was made. From this pooled sample, three samples were derived, which we called Technical Replicates and mRNA-seq was performed.

Overall our experimental design consisted of two genotypes and two conditions. In each case three technical replicates were used. At the end we have total 12 samples (2 genotypes X 2 treatment (well-watered vs drought) X 3 technical replicates). We are interested in finding differentially expressed genes in each genotype with respect to treatment (i.e. Genotype 1: Well-watered vs Drought; Genotype 2: Well-watered vs Drought).

My question is can I use technical replicates to fit the model in Deseq2? Is it appropriate? if not suitable, what are the alternatives to analyse such a data?

Thank you very much. Best, Ahmad

It's technically possible and not too uncommon since often one does not have enough RNA from individual specimen to make a library. It's not a technical replicate as you're not running replication on the same pool of RNA or sequencing the same library several times. You're pooling individuals so it is still some sort of a biological replicate, as it measures pools of individuals between groups. Not perfect but you anyway don't seem to have a choice. Just go along with it.