makeDataObject: Makes data object for fitting the random effects model

In Sage-Bionetworks/snm: Supervised Normalization of Microarrays

Description

Makes data object for fitting the random effects model

Usage

makeDataObject(Y.pooled, np, snm.obj, exp, bins)

Arguments

Y.pooled	Matrix of pooled raw data. Element i,j is the average raw data for every probe in bin i on array j.
np	list matching bins to probes
snm.obj	An object of class snm
exp	New environment to hold the lmer formula
bins	list matching bins to probes

Author(s)

Brig Mecham <[e-mail]> and John D. Storey <[e-mail]>

Examples

##---- Should be DIRECTLY executable !! ----
##-- ==>  Define data, use random,
##--	or do  help(data=index)  for the standard data sets.

## The function is currently defined as
function (Y.pooled, np, snm.obj, bSM.model, exp,bins) 
{
   lnp <- length(np)
   num.arrays <- dim(Y.pooled)[2]
   signal <- data.frame(y = as.numeric(Y.pooled), probes = factor(rep(1:lnp, 
                                             times = num.arrays)))
   f <- lapply(1:length(snm.obj$int.var), function(x) {
     i <- snm.obj$int.var[, x]
     rep(i, each = lnp)
   })
   f <- as.data.frame(f)
   colnames(f) <- names(snm.obj$int.var)
   if (dim(snm.obj$adj.var)[2] > 1) {
     z <- lapply(2:(dim(snm.obj$adj.var)[2]), function(x) {
       i <- snm.obj$adj.var[, x]
       rep(i, each = lnp)
     })
     z <- as.data.frame(z)
     colnames(z) <- colnames(snm.obj$adj.var)[-1]
     df <- cbind(cbind(cbind(signal, z), f), bSM.model)
   }else {
     df <- cbind(cbind(signal, f), bSM.model)
   }
   df$weights = sapply(bins,length) / 50 #mean(sapply(bins,length))
   df$weights[df$weights >1] <- 1
   TMP <- sapply(1:dim(df)[2], function(x) {
     assign(colnames(df)[x], df[, x], envir = exp)
   })
   as.data.frame(df)
 }

Sage-Bionetworks/snm documentation built on May 9, 2019, 12:14 p.m.