IMR Press / FBL / Volume 9 / Issue 1 / DOI: 10.2741/1186

Frontiers in Bioscience-Landmark (FBL) is published by IMR Press from Volume 26 Issue 5 (2021). Previous articles were published by another publisher on a subscription basis, and they are hosted by IMR Press on as a courtesy and upon agreement with Frontiers in Bioscience.

Statistical methods for detecting genomic alterations through array-based comparative genomic hybridization (CGH)
Show Less
1 Department of Statistics and Applied Probability, University of California, Santa Barbara, California 93106, USA
2 Department of Pediatrics and Division of Biostatistics, Medical College of Wisconsin, 8701 Watertown Plank Road, Milwaukee, WI 53226, USA
Front. Biosci. (Landmark Ed) 2004, 9(1), 540–549;
Published: 1 January 2004

Array-based comparative genomic hybridization (ABCGH) is an emerging high-resolution and high-throughput molecular genetic technique that allows genome-wide screening for chromosome alterations associated with tumorigenesis. Like the cDNA microarrays, ABCGH uses two differentially labeled test and reference DNAs which are cohybridized to cloned genomic fragments immobilized on glass slides. The hybridized DNAs are then detected in two different fluorochromes, and the significant deviation from unity in the ratios of the digitized intensity values is indicative of copy-number differences between the test and reference genomes. Proper statistical analyses need to account for many sources of variation besides genuine differences between the two genomes. In particular, spatial correlations, the variable nature of the ratio variance and non-Normal distribution call for careful statistical modeling. We propose two new statistics, the standard t-statistic and its modification with variances smoothed along the genome, and two tests for each statistic, the standard t-test and a test based on the hybrid adaptive spline (HAS). Simulations indicate that the smoothed t-statistic always improves the performance over the standard t-statistic. The t-tests are more powerful in detecting isolated alterations while those based on HAS are more powerful in detecting a cluster of alterations. We apply the proposed methods to the identification of genomic alterations in endometrium in women with endometriosis.

Array-based comparative genomic hybridization
genomic alterations
hybrid adaptive splines
statistical methods
Back to top