Using model-based variance estimates instead of bootstraps

[dmoaks]

Hello,

I am trying to save computational time and use "modelbased" as the "var_estimate" parameter of the gseaAfterBoot function. However, the function is not working because it seems to require the "boots" input even when it doesn't use it. Can you post example code of how to properly use the "modelbased" parameter?

[amcdavid]

At the moment, you need to pass in 2 bootstrap replicates. These are just used to get the dimensions of some arrays that hold the result. This is not intrinsically necessary, and can/should be fixed.

 data(vbetaFA)
 vb1 = subset(vbetaFA, ncells==1)
 vb1 = vb1[,freq(vb1)>.1][1:15,]
 zf = zlm(~Stim.Condition, vb1)
 boots = bootVcov1(zf, 2)
 sets=list(A=1:5, B=3:10, C=15, D=1:5)
 gsea=gseaAfterBoot(zf, boots, sets, CoefficientHypothesis('Stim.ConditionUnstim'), control = list(var_estimate = 'modelbased', n_randomize = Inf))

[amcdavid]

Closed in 04e17c31f0b28d1503d197b51846026e403835d2