Journal Article

Copy number variation has little impact on bead-array-based measures of DNA methylation

E. Andrés Houseman, Brock C. Christensen, Margaret R. Karagas, Margaret R. Wrensch, Heather H. Nelson, Joseph L. Wiemels, Shichun Zheng, John K. Wiencke, Karl T. Kelsey and Carmen J. Marsit

in Bioinformatics

Volume 25, issue 16, pages 1999-2005
Published in print August 2009 | ISSN: 1367-4803
Published online June 2009 | e-ISSN: 1460-2059 | DOI: http://dx.doi.org/10.1093/bioinformatics/btp364
Copy number variation has little impact on bead-array-based measures of DNA methylation

More Like This

Show all results sharing this subject:

  • Bioinformatics and Computational Biology

GO

Show Summary Details

Preview

Motivation: Integration of various genome-scale measures of molecular alterations is of great interest to researchers aiming to better define disease processes or identify novel targets with clinical utility. Particularly important in cancer are measures of gene copy number DNA methylation. However, copy number variation may bias the measurement of DNA methylation. To investigate possible bias, we analyzed integrated data obtained from 19 head and neck squamous cell carcinoma (HNSCC) tumors and 23 mesothelioma tumors.

Results: Statistical analysis of observational data produced results consistent with those anticipated from theoretical mathematical properties. Average beta value reported by Illumina GoldenGate (a bead-array platform) was significantly smaller than a similar measure constructed from the ratio of average dye intensities. Among CpGs that had only small variations in measured methylation across tumors (filtering out clearly biological methylation signatures), there were no systematic copy number effects on methylation for three and more than four copies; however, one copy led to small systematic negative effects, and no copies led to substantial significant negative effects.

Conclusions: Since mathematical considerations suggest little bias in methylation assayed using bead-arrays, the consistency of observational data with anticipated properties suggests little bias. However, further analysis of systematic copy number effects across CpGs suggest that though there may be little bias when there are copy number gains, small biases may result when one allele is lost, and substantial biases when both alleles are lost. These results suggest that further integration of these measures can be useful for characterizing the biological relationships between these somatic events.

Contact: E_Andres_Houseman@brown.edu

Supplementary information: Supplementary data are available at Bioinformatics online.

Journal Article.  5016 words.  Illustrated.

Subjects: Bioinformatics and Computational Biology

Full text: subscription required

How to subscribe Recommend to my Librarian

Users without a subscription are not able to see the full content. Please, subscribe or login to access all content.