Supplementary Material for "Statistical issues in the analysis of Illumina data"

Mark Dunning, 7th November 2007

Raw Data

Raw data for each of the eight BeadChips can obtained via the following FTP links. Each zip file is around 750MB in size and contains the TIFF image and text file for each of the six arrays on the chip. Spike concentrations 1000pM to 3pM can be found on chips 1377192001-8 and concentrations 1pM to 0pM are found on chips 1377192013-20. Once unzipped, these raw data can be read using the 'readIllumina' function in beadarray (see the R script below for specific commands).

Use the username 'illumina' and the password 'spike' to login to the FTP site

Diagnostic plots

Bead level data

Image plots of foreground intensities for each strip on each BeadChip (log-2 transformed)
Image plots of background intensities for each strip on each BeadChip (log-2 transformed)
Boxplots of foreground intensities (background subtracted, log-2 transformed and colour coded by BeadChip)
Boxplots of background intensities (log-2 transformed and colour coded by BeadChip)

Bead summary data

Smoothed scatter plots of summarised intensities (log-2) of individual arrays against median array
Boxplots of summarised intensities per array (colour coded by BeadChip, non-normalised)
Boxplots of standard errors per array (colour coded by BeadChip, non-normalised)
Boxplots of number of beads per bead type per array (colour coded by BeadChip, non-normalised)
Number of false discoveries versus number of genes selected for each contrast. Bead types were ranked using log-odds.

Additional Files

Targets file Text file listing all array names and spike concentrations
Results of the re-annotation for all non-spikes and non-controls
Intensities of each control probe on every array as exported by BeadStudio
Averaged control profile as exported from BeadStudio
Sequences for the Illumina controls used in the experiment
Pre-calculated thermodynamic values for each spike probe
File giving bead IDs, targets and sequences for the spikes
Bead summary data .rda (sharpened, locally background corrected, summarised)
BeadStudio output (non-normalised). The file SampleProbeProfile.txt contains summary data for the non-spikes. The file ControlProbeProfile.txt contains summary data for the spikes and other control probes. The spikes have 'DRC' as their TargetIDs.

We are grateful to Illumina for allowing us to make this data set publicly available. In particular, we thank Gary Nunn and Semyon Kruglyak for retrieving these data for us.

Scripts

Full Analysis Script R code to read the bead level data above and re-produce all the figures and numerical values quoted in the paper. Note that reading the bead level data requires around 4Gb of RAM.