Quality Control of Microarrays

Once you have your raw data it is important to look at some array quality metrics. These will tell you if there are any problems with the array experiments, such as problems with the arrays themselves or with sample degredation. Each microarray manufacturer has their own set of built in tests, which are described in the array documentation.

To use the Bioconductor packages for assessing Microarray quality you need to install the ArrayQualityMetrics package.

biocLite("arrayQualityMetrics")
library(arrayQualityMetrics)
arrayQualityMetrics(expressionset = normalized_rma, outdir = "QCnorm", spatial = FALSE,
+ force = TRUE, intgroup = "Type", grouprep = TRUE)


This is limited by memory and that is why I have limited the number of CEL files in the code to the first 20. The processing will then take a while to run all the diagnostics. It worked in verions 2.10 but there are problems with R 2.12 and Bioconductor 2.7 that it does not automatically load the libraries.

There are also quality control measures for Affymetrix data in the simpleaffy package. This is run with the function qc().

library("simpleaffy")
rawqc <- qc(raw[,1:20])
plot (rawqc, cex=0.5)



This shows that there is an appreciable amount of RNA degredation on many of the arrays and this is affecting the 3 prime to 5 prime ratio of actin. Many of the actin measures are actually outliers. This might suggest problems with the experimental protocol for setting up the arrays.

Another alternative is the affyQCReport package, which produces a pdf output of the quality control data.

library("arrayQualityMetrics")
QCReport(raw[,1:20])