I am designing a RNA-seq experiment to study transcriptomic profile changes upon parasite exposure. Some papers I read, mostly extract RNA from the entire treatment group (organisms exposed to parasite), without (seemingly) bothering if all the organisms were indeed affected. Doesn't this practice bias results? Are there any studies on how this could affect results?
It will dilute your signal if you have a pool of samples +/- the parasite prepared as one library prep. It is always better to maximize the number sequencing libraries (usually cost prohibitive) and extract RNA from samples you are confident exemplify the condition of interest, and an equal number of control samples to maximize your statistical power.