Open Access Research article

Relative impact of key sources of systematic noise in Affymetrix and Illumina gene-expression microarray experiments

Robert R Kitchen123, Vicky S Sabine4, Arthur A Simen3, J Michael Dixon5, John MS Bartlett4 and Andrew H Sims1*

Author Affiliations

1 Applied Bioinformatics of Cancer Group, Breakthrough Breast Cancer Research Unit, Institute of Genetics and Molecular Medicine, Crewe Road South, Edinburgh, Edinburgh, EH4 2XR, UK

2 School of Physics, University of Edinburgh, 10 Crichton Street, Edinburgh, EH8 9AB, UK

3 Yale University School of Medicine, Department of Psychiatry, 300 George Street, Suite 901, New Haven, CT 06511, USA

4 Endocrine Cancer Group, Edinburgh Cancer Research Centre, Institute of Genetics and Molecular Medicine, Crewe Road South, Edinburgh, EH4 2XR, UK

5 Breast Cancer Research Group, Western General Hospital, Crewe Road South, Edinburgh, EH4 2XU, UK

For all author emails, please log on.

BMC Genomics 2011, 12:589  doi:10.1186/1471-2164-12-589

Published: 1 December 2011

Additional files

Additional file 1:

Supplementary material S1. Comparison of estimated variance components in Affymetrix and Illumina data with and without background correction as part of the array pre-processing.

Format: PDF Size: 1.5MB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 2:

Supplementary material S2. Illustration of our Illumina Ref-8 (experiment 1) and HT-12 (experiments 2 & 3) BeadChips, processed in eight batches (also referred to as 'runs') corresponding to the different days on which the samples were hybridised and scanned. UHRR samples are labelled as C1-25. Replicate breast tumour clinical samples are identified with a suffix of 'a' through 'd'. The pre- and post-treatment biopsy samples are identified by a triangle to the left and right of the sample IDs, respectively.

Format: PDF Size: 38KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 3:

Supplementary material S3. Correlation of expression change as a result of inter-run and inter-chip technical variation between UHRR and pooled controls with tumour duplicates. Tumour duplicates (individually plotted) are arranged on the x-axis to be close to others processed on the same BeadChip. UHRR and both types of pooled-control (comprised of pre- and post-treatment tumour RNA, respectively) are correlated more strongly with individual tumour duplicates in which one 'half' of the duplicate was processed on the outlying chip 22. Pooled-controls also consistently score slightly higher correlation than UHRR.

Format: PDF Size: 157KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 4:

Supplementary material S4. Scatter plots of fold-changes between tumour duplicates and replicate pools. These plots each show the magnitude of the change in expression between technical replicates introduced by the runs and show how well correlated such changes are between the Pool/tumour-duplicates. In the figure, the pre-treatment tumour duplicate are coloured blue and post-treatment are green; probes that are differentially regulated, up or down, at least two-fold due to the different runs are highlighted in each plot (red points). Note that most of the samples with a duplicate on chip 22 are subject to a greater magnitude of variation than samples with a duplicate on chip 21.

Format: PNG Size: 928KB Download file

Open Data

Additional file 5:

Supplementary material S5. Plot of GC-content vs. (1/probe signal).

Format: PDF Size: 207KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 6:

Supplementary material S6. Distribution of MAQC Illumina probes as a fraction of target gene length. Light blue points are used for probes that mapped to the antisense DNA strand and dark blue for the sense strand. Left plot represents the fraction of target gene length in terms of 3' and 5' coordinates on each strand while the right plot is 'normalised' for the anti-sense strand and plots the fraction in terms of absolute position along the DNA molecule.

Format: PDF Size: 549KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 7:

Supplementary material S7. Probe position against probe standard deviation: Plots of probe position against probe standard deviation estimated at the inter-laboratory (A), inter-chip (B), and inter-array (C) levels in the MAQC Illumina dataset. Light and dark blue points again identify probes that mapped to the antisense and sense strands, respectively.

Format: PDF Size: 2.8MB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 8:

Supplementary material S8. Quality control metrics over all of our Affymetrix and Illumina arrays including array-level intensity distributions, Illumina bead-standard errors, and bead-representation distributions between the two Illumina array-versions.

Format: PDF Size: 2.8MB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data