I have been spending a lot of time looking at scRNA seq and am now looking at some bulk data and have a few questions. Right now I am attempting to determine if a subset of known genes (32) are differentially expressed across two samples. I am not sure how to go about this considering I essentially have an n of 1 for both treatments.
Is there a way that this is usually done? Currently, since the log fold change of all genes seems normally distributed and centered around zero I am guessing that I can claim that the two samples are similar, and I see from the histogram that most genes are expressed at similar levels. From there I could get a standard deviation or something and report how far from the center these genes fall, but that does not seem satisfying because the variation is coming from within genes rather than between samples.
No, not without replicates. Please use google and the search function for experiments without replicates to learn why.
There are plenty of related questions out there. Please go through their answers.
Currently, since the log fold change of all genes seems normally
distributed and centered around zero I am (...)
Again, without statistics FCs are meaningless. I do not think this statement has a solid ground.
From there I could get a standard deviation or something and report
how far from the center these genes fall, but that does not seem
satisfying because the variation is coming from within genes rather
than between samples.
Again, claims about similarities and significantes need replicates. Nothing you can do in silico do rescue a suboptimally-designed experiment.
