Fitted values from EdgeR
Last seen 5 months ago
I wonder of the values in lrt$fitted.values are cpm normalized values?
des <- model.matrix(~KRAS)
fit <- glmFit(y2, des)
lrt <- edgeR::glmLRT(fit)
Last seen 3 hours ago
glmLRT> d <- DGEList(y)
glmLRT> # Likelihood ratio tests for trend
glmLRT> results <- glmLRT(fit, coef=2)
x0 x1 x2
gene.1 9.618113 29.449462 69.696630
gene.2 4.787029 5.444920 4.787005
gene.3 8.071594 7.879561 5.945545
gene.4 5.807543 6.605920 5.807928
gene.5 6.194186 5.686449 4.035013
gene.6 3.528931 5.335160 6.234458
x0 x1 x2
gene.1 6 44 55
gene.2 5 5 5
gene.3 7 10 5
gene.4 8 2 8
gene.5 5 8 3
gene.6 5 2 8
Does that answer your question?
Last seen 5 hours ago
WEHI, Melbourne, Australia
The help page for glmLRT explains the meaning of all the output components.
For more mathematical detail, here's the paper that describes the methodology:
The quantities that are called fitted.values in the output are denoted by the Greek letter mu in the article. Generally speaking, these quantities are for internal edgeR use and you do not need to manipulate them yourself in a practical analysis.
Traffic: 129 users visited in the last hour