This function plots the dependence of mean noise per cell on the total UMI count per cell. It serves as a basis for choosing the prior parameter gamma
(see function compTBNoise
). With a proper parameter choice, there should be no correlation between the two quantities. If a positive correlation is observed, gamma
should be increased in order to weaken the prior. If the correlation is negative, gamma
should be decreased in order to increase the strength of the prior.
plotUMINoise(object, noise, log.scale = TRUE)
RaceID SCseq
object.
object returned by compTBNoise
function.
Logical. If TRUE
total transcript counts and epsilon
estimates are log2-transformed for plotting. Default is TRUE
.