RobNorm is a R package to robustly normalize an expression matrix, analyzed in the paper "RobNorm: Model-Based Robust Normalization for High-Throughput Proteomics from Mass Spectrometry Platform".
Please contact Meng Wang by email mengw1@stanford.edu for questions.
library(devtools)
install_github("mwgrassgreen/RobNorm")
library(RobNorm)
norm.result = RobNorm(X.0, gamma.0=0.5, tol=10^(-4), step=200)
To simulate an expression matrix
sim.result = sim.dat.fn(row.frac=0.2, col.frac=0.2, mu.up=3, mu.down=-3, n=5000, m=200, nu.fix=TRUE)
X.0 = sim.result$dat
norm.result = RobNorm(X.0, gamma.0=0.5)
X.0.norm = norm.result$norm.data
To compare sample boxplots before and after normalization
par(mfrow=c(2,1))
boxplot(X.0, main="Sample boxplots before normalization", ylab="expression", xlab="sample", cex.main=1.5, cex.lab=1.5)
boxplot(X.0.norm, main="Sample boxplots after normalization", ylab="expression", xlab="sample", cex.main=1.5, cex.lab=1.5)
Since in the simulation nu.fix=TRUE meaning the underlying nu = 0, the boxplots before and after normalization.
After the normalization step, to check our AdaTiSS package for calcualizating robust z-score for each condition.
For more statistical analysis, to check our AdaReg package.