Last active
February 14, 2016 05:08
-
-
Save maromato/0c941b3f68154ac6c460 to your computer and use it in GitHub Desktop.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
#Initially, you need to normalize raw microarray data and make a spread sheet for gene expression as shown elsewhere. | |
#input spread sheet for microarray data | |
x = read.table("XXXX.txt",header=T,sep="\t") | |
x2 = x[,2:ncol(x)] | |
x2 = as.matrix (x2) | |
rownames(x2) | |
#statistical analysis for differential expression | |
library(limma) | |
design = cbind(Cell_A = c(1,1,1,0,0,0),Cell_B = c(0,0,0,1,1,1)) | |
fit = lmFit(x2, design=design)# Fit the original matrix to the above design. | |
contrastsMatrix = makeContrasts("Cell_B-Cell_A", | |
levels = design)# We want to compare A vs. B, A vs. C and B vs. C | |
fit2 = contrasts.fit(fit, contrasts = contrastsMatrix) # Making the comparisons. | |
out = eBayes(fit2) # Moderating the t-tetst by eBayes method. | |
p.value = out$p.value #to put p.values for indevisual genes into the vector, p.value | |
q.value = apply(p.value, MARGIN=2, p.adjust, method="BH")#to put | |
q.values for indevisual genes into the vector, a.value | |
ranking = apply(p.value, MARGIN=2, rank)#to put the ranking for indevisual genes in term of p.value into the vector, ranking | |
tmp = cbind(x2, p.value, q.value, ranking) | |
#extract mouse TF data from original spread sheet. | |
b = read.table("mouseTF.txt",sep=" \t",header=F)#you need to make a entire list of the mouse transcription factors from the information | |
in the site (http://genome.gsc.riken.jp/TFdb/tf_list.html) | |
obj = rownames(x2) %in% b[,1] # to find the rows that are included in the list of transcription factors (https://www.biostars.org/p/5737/) | |
tmp2 = tmp[obj,] | |
ob = rank(tmp2[,ncol(tmp2)-2])< 65 | |
tmp3 = tmp2[ob,] | |
z = tmp3[,1:6] | |
#Clustering and visualization. | |
library(gplots) | |
heatmap.2(as.matrix(z), col=greenred(75), scale="row", key=T, keysize=1.5,density.info="none", trace="none",cexCol=0.9, cexRow=0.5) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment