R Lab - Part I

A Cancer Modeling Example

See StatPrinciples_RLab.pdf for some background info.

Exercise on analysis of miRNA, mRNA and protein data from the paper Aure et al, Integrated analysis reveals microRNA networks coordinately expressed with key proteins in breast cancer, Genome Medicine, 2015.

Please run the code provided to replicate some of the analyses in Aure et al. (2015). Make sure you can explain what all the analysis steps do and that you understand all the results.

In addition, there are three extra tasks Task 1, Task 2, Task 3, where no R code is provided. Please do these tasks when you have time available at the end of the lab.

Install R packages

install.packages("glmnet")
install.packages("gclus")

Load the data

Read the data, and convert to matrix format.

mir <- read.table("lab/data/miRNA-421x282.txt", header=T, sep="\t", dec=".")
rna <- read.table("lab/data/mRNA-100x282.txt", header=T, sep="\t", dec=".")
prt <- read.table("lab/data/prot-100x282.txt", header=T, sep="\t", dec=".")

# Convert to matrix format

mir <- as.matrix(mir)
rna <- as.matrix(rna)
prt <- as.matrix(prt)

Print the data

mir[1:4, 1:4]

            OSL2R.3002T4 OSL2R.3005T1 OSL2R.3013T1 OSL2R.3030T2
hsa-let-7a      -1.10330      0.40033     -0.65364      0.78277
hsa-let-7a*     -0.58049     -0.72246      1.46023     -0.67980
hsa-let-7b      -3.17134      0.41602     -0.13054      1.11872
hsa-let-7c      -3.10923      0.27861     -0.17365      0.47395

rna[1:4, 1:4]

      OSL2R.3002T4 OSL2R.3005T1 OSL2R.3013T1 OSL2R.3030T2
ACACA      1.60034     -0.49087     -0.26553     -0.27857
ANXA1     -2.42501     -0.05416     -0.46478     -2.18393
AR         0.39615     -0.43348     -0.10232      0.58299
BAK1       0.78627      0.39897      0.22598     -1.31202

prt[1:4, 1:4]

      OSL2R.3002T4 OSL2R.3005T1 OSL2R.3013T1 OSL2R.3030T2
ACACA      0.48181     -0.76244      0.22878      0.02493
ANXA1     -0.37323      0.52558      0.73313     -1.60107
AR         1.39394     -0.33711      0.07152      1.51886
BAK1       1.44828      1.26768     -0.32807      1.41299

Visualise the overall distribution of expression levels by histogram

par(mfrow=c(3,1))
hist(mir, nclass=40, xlim=c(-5,5), col="lightblue")
hist(rna, nclass=40, xlim=c(-5,5), col="lightblue")
hist(prt, nclass=40, xlim=c(-5,5), col="lightblue")

mRNA-protein associations (only first nine genes)

par(mfrow=c(3,3))
par(mar=c(3,3,3,3))
for (i in 1:9) {
  plot(rna[i,], prt[i,], pch=19)
  title(rownames(rna)[i])
  lines(smooth.spline(rna[i,], prt[i,], df=4), col="red")
}

Task 1

Extend the above analysis to cover all genes.

Explore the correlations

Compute and plot mRNA-protein correlations

rho = rep(NA, nrow(rna))  
for (i in 1:nrow(rna)) {
  rho[i] = cor(rna[i,], prt[i,])
}
par(mfrow=c(1,1))
hist(rho, col="lightblue")

Calculate the correlation of each miRNA to each protein

RHO = matrix(NA, nrow(mir), nrow(prt))
for (i in 1:nrow(mir)) {
  for (j in 1:nrow(prt)) {
     RHO[i,j] = cor(mir[i,], prt[j,]) 
  }
}
par(mfrow=c(1,1))
hist(RHO, col="lightblue")

Visualize as heatmap

Use the code provided to visualize the heatmap.

Note: If you get the error message Error in plot.new() : figure margins too large, try to increase the Plots panel in your RStudio workspace.

source("lab/code/clustermap_beta.R")

plot.init(tree=c(2,3))
hcluster(RHO, clust="row", distance="euclidean", linkage="complete")
hcluster(RHO, clust="col", distance="euclidean", linkage="complete")
plot.hmap(RHO)
plot.tree(side=2)
plot.tree(side=3)
plot.hmap.key()

Task 2

Compare this heatmap with Figure 3 in Aure et al. (2015). Are these two figures showing the same results?

--- title: "R Lab - Part I" format: html: code-fold: false code-tools: true --- ## A Cancer Modeling Example See `StatPrinciples_RLab.pdf` for some background info. Exercise on analysis of miRNA, mRNA and protein data from the paper Aure et al, Integrated analysis reveals microRNA networks coordinately expressed with key proteins in breast cancer, Genome Medicine, 2015. Please run the code provided to replicate some of the analyses in Aure et al. (2015). Make sure you can explain what all the analysis steps do and that you understand all the results. In addition, there are three extra tasks `Task 1`, `Task 2`, `Task 3`, where no R code is provided. Please do these tasks when you have time available at the end of the lab. ### Install R packages ```{r} #| label: install-packages #| warning: false #| echo: true #| eval: false install.packages("glmnet") install.packages("gclus") ``` ### Load the data Read the data, and convert to matrix format. ```{r} #| label: load-data #| warning: false #| echo: true mir <- read.table("lab/data/miRNA-421x282.txt", header=T, sep="\t", dec=".") rna <- read.table("lab/data/mRNA-100x282.txt", header=T, sep="\t", dec=".") prt <- read.table("lab/data/prot-100x282.txt", header=T, sep="\t", dec=".") # Convert to matrix format mir <- as.matrix(mir) rna <- as.matrix(rna) prt <- as.matrix(prt) ``` Print the data ```{r} #| label: look-at-data #| warning: false #| echo: true mir[1:4, 1:4] rna[1:4, 1:4] prt[1:4, 1:4] ``` Visualise the overall distribution of expression levels by histogram ```{r} par(mfrow=c(3,1)) hist(mir, nclass=40, xlim=c(-5,5), col="lightblue") hist(rna, nclass=40, xlim=c(-5,5), col="lightblue") hist(prt, nclass=40, xlim=c(-5,5), col="lightblue") ``` mRNA-protein associations (only first nine genes) ```{r} par(mfrow=c(3,3)) par(mar=c(3,3,3,3)) for (i in 1:9) { plot(rna[i,], prt[i,], pch=19) title(rownames(rna)[i]) lines(smooth.spline(rna[i,], prt[i,], df=4), col="red") } ``` ::: callout-note ## Task 1 Extend the above analysis to cover all genes. ::: ## Explore the correlations Compute and plot mRNA-protein correlations ```{r} rho = rep(NA, nrow(rna)) for (i in 1:nrow(rna)) { rho[i] = cor(rna[i,], prt[i,]) } par(mfrow=c(1,1)) hist(rho, col="lightblue") ``` Calculate the correlation of each miRNA to each protein ```{r} RHO = matrix(NA, nrow(mir), nrow(prt)) for (i in 1:nrow(mir)) { for (j in 1:nrow(prt)) { RHO[i,j] = cor(mir[i,], prt[j,]) } } par(mfrow=c(1,1)) hist(RHO, col="lightblue") ``` ## Visualize as heatmap Use the code provided to visualize the heatmap. **Note:** If you get the error message `Error in plot.new() : figure margins too large`, try to increase the *Plots* panel in your RStudio workspace. ```{r} #| echo: true #| eval: false source("lab/code/clustermap_beta.R") plot.init(tree=c(2,3)) hcluster(RHO, clust="row", distance="euclidean", linkage="complete") hcluster(RHO, clust="col", distance="euclidean", linkage="complete") plot.hmap(RHO) plot.tree(side=2) plot.tree(side=3) plot.hmap.key() ``` ::: callout-note ## Task 2 Compare this heatmap with Figure 3 in Aure et al. (2015). Are these two figures showing the same results? :::