Confident non-linear effect sizes using limma

limma_nonlinear_confects(object, design, effect, fdr = 0.05, step = 0.01,
  trend = FALSE)

Arguments

object

An expression matrix or EList object, or anything limma's lmFit will accept.

design

Design matrix.

effect

A non-linear effect, created with one of the effect_... functions.

fdr

False Discovery Rate to control for.

step

Granularity of log2 fold changes to test.

trend

Should eBayes(trend=TRUE) be used?

Value

Technical note: Signed confects are based on TREAT-style p-values. Unsigned confects (generally with df>1) are based on comparing the best fit within the H0 region to the best fit overall, which may up to double p-values.

See nest_confects for details of how to interpret the result.