Mean-variance trend in voom

Question

[hi.jia.wu]

I'm performing voom RNAseq analysis on GSE86507. My mean-variance trend of all my samples seem to have an unusual trend with some systematic dispersion on the lower ends of the means (there are aggressive piecewise linear trends extending like rays). Has anyone seen this trend before and have some insight as to what may be the cause of it? Link to image: https://imgur.com/a/SR06aHc

[Ahill]

Rays on these types of plots extending from very low counts can sometimes be due to discretization of small integer counts (e.g, 0.5+0,1,2,3, ...) and variances particularly if n is small. How many replicates do you have included the mean/variance estimates? You could grab the source data from the mean/variance plot under some of the rays (maybe using R identify()) to confirm or deny if that is occurring.

[Kristoffer Vitting-Seerup]

Is it single cell data?