Labs for Advanced Stats

Clusters from Expression Data - Lab Exercises

This lab analyzes gene expression data from a breast cancer study. There are 24 samples with 12625 expression measures. Here is a way to read the data in.

```{r data}
#source("https://bioconductor.org/biocLite.R")
#biocLite()
if (!require("Biobase")) install.packages("Biobase")
library(Biobase)
#download.file("ftp://ftp.umiacs.umd.edu/pub/hcorrada/chang03.rda", destfile="./chang03.rda")
load("chang03.rda")
show(chang03)
```

I'd recommend doing a couple of things before you get too deeply into it ...

```{r data frame}
descrDF = pData(chang03)
names(descrDF)
table(descrDF$disease.state,descrDF$Progesterone..receptor.status)
exprDF = exprs(chang03)
dim(exprDF)
summary(exprDF)
boxplot(exprs(chang03))
exprDF_log <- log2(exprDF)
boxplot(exprDF_log,col=as.numeric(descrDF$disease.state)+1)
```

Make friends with the data. Explore the dependent and independent variables.
Calculate a distance matrix between the samples
Perform a cluster analysis based on these distances
Visualize them as a cluster plot
Perform additional analyses that allow you to make sense of the expression differences among the samples

Competences earned this week:

Examine the multivariate distances of a genomic data set
Relate multivariate distances to a grouping variable

last modified: 9/30/17