Hello. I'm analyzing a large plant gene atlas study and we're seeing some strange behavior for a set of samples (reps) for a particular tissue. This tissue is known to have a smaller overall number of expressed genes versus other tissues with more highly expressed genes. Plots below are the library sizes before and after normalization (with the 4 unusual samples denoted in red):
And here are boxplots of the counts per gene per sample before and after normalization (the unusual 4 samples are denoted in blue here)
We do NOT want to do differential expression analysis on these data. We just want to generate plots to visualize the normalized gene atlas data. Typical plant gene atlas studies that I've seen published just use RPKM normalization and plot the RPKM values. I'd prefer to use a more sophisticated form of normalization given all of the problems with RPKM normalization. Any advice on how to handle this data would be appreciated.