SAHA and Depsipeptide HDACi:
Time course experiment with 2 treatments

1. Aims

This exercise uses advanced linear modeling and empirical Bayes moderated F-statistics to identify genes with interesting patterns in a multiple time course experiment.

3. The experiment

This laboratory uses data from Peart et all (2005) from 36 cDNA microarrays. Histone deacetylase inhibitors (HDACi) are compounds which inhibit tumor cell growth and survival. They are currently in early phase clinical trials as a cancer treatment. The structurally diverse HDACi SAHA and depsipeptide activate a number of common molecular pathways to induce apoptosis, however there are subtle differences in their mechanisms of action. Here we focus on finding genes which are specific to each of the two compounds, SAHA and depsipeptide.

4. Load the data

> setwd("sahadepsi")
> library(limma)
> load("MA.RData")
> objects()
> MA$targets

3. Estimate the correlation within blocks

> block <- rep(1:6,each=6)
> fitcor <- duplicateCorrelation(MA,design,block=block)
Loading required package: statmod
> fitcor$consensus
[1] 0.5107596

4. Fit the linear model

> fit <- lmFit(MA,design,block=block,correlation=fitcor$consensus)
> boxplot(fit$coef~col(fit$coef),names=colnames(fit$coef),las=2,ylab="Log2 Fold Change")

5. Extract time trends and interactions

Time trends in response to SAHA:

> cont.matrix <- cbind(
+ S1=c(-1,1,0,0,0,0,0,0,0,0,0,0),
+ S2=c(-1,0,1,0,0,0,0,0,0,0,0,0),
+ S4=c(-1,0,0,1,0,0,0,0,0,0,0,0),
+ S8=c(-1,0,0,0,1,0,0,0,0,0,0,0),
+ S16=c(-1,0,0,0,0,1,0,0,0,0,0,0)
+ )
> fits <- contrasts.fit(fit, cont.matrix)
> fits <- eBayes(fits)

Time trends in response to depsipeptide:

> cont.matrix <- cbind(
+ D1=c(0,0,0,0,0,0,-1,1,0,0,0,0),
+ D2=c(0,0,0,0,0,0,-1,0,1,0,0,0),
+ D4=c(0,0,0,0,0,0,-1,0,0,1,0,0),
+ D8=c(0,0,0,0,0,0,-1,0,0,0,1,0),
+ D16=c(0,0,0,0,0,0,-1,0,0,0,0,1)
+ )
> fitd <- contrasts.fit(fit, cont.matrix)
> fitd <- eBayes(fitd)

Interaction effects:

> cont.matrix <- cbind(
+ I1=c(1,-1,0,0,0,0,-1,1,0,0,0,0),
+ I2=c(1,0,-1,0,0,0,-1,0,1,0,0,0),
+ I4=c(1,0,0,-1,0,0,-1,0,0,1,0,0),
+ I8=c(1,0,0,0,-1,0,-1,0,0,0,1,0),
+ I16=c(1,0,0,0,0,-1,-1,0,0,0,0,1)
+ )
> fiti <- contrasts.fit(fit, cont.matrix)
> fiti <- eBayes(fiti)

Summary:

> results <- new("TestResults",matrix(0,nrow(fiti),3))
> results[,1] <- p.adjust(fits$F.p.value,method="fdr") < 0.05
> results[,2] <- p.adjust(fitd$F.p.value,method="fdr") < 0.05
> results[,3] <- p.adjust(fiti$F.p.value,method="fdr") < 0.05
> colnames(results) <- c("SAHA","Depsi","Difference")
> vennDiagram(results)

6. SAHA and depsipeptide specific genes

Find genes responding only to SAHA, and only to depsipeptide:

> SAHAOnly <- results[,"SAHA"]==1 & results[,"Depsi"]==0 & results[,"Difference"]==1
> DepsiOnly <- results[,"SAHA"]==0 & results[,"Depsi"]==1 & results[,"Difference"]==1
 
> par(mfrow=c(1,2))
> plotlines(fits$coef[SAHAOnly,],ylim=c(-1.5,1.5),ylab="lfc",main="Response to SAHA")
> plotlines(fitd$coef[SAHAOnly,],ylim=c(-1.5,1.5),ylab="lfc",main="Response to Depsi")

References

1. Peart, M. J., Smyth, G. K., van Laar, R. K., Richon, V. M., Holloway, A. J., Johnstone, R. W. (2005). Identification and functional significance of genes regulated by structurally diverse histone deacetylase inhibitors. Proceedings of the National Academy of Sciences of the United States of America.102, 3697-3702. [Publisher Full Text]
2. Smyth, G. K. (2004). Linear models and empirical Bayes methods for assessing differential expression in microarray experiments. Statistical Applications in Genetics and Molecular Biology 3, No. 1, Article 3. [Publisher Full Text] (Introduces moderated F-statistics)

Return to list of Exercises