[Show abstract][Hide abstract] ABSTRACT: Comparative genome hybridization (CGH) to DNA microarrays (array CGH) is a technique capable of detecting deletions and duplications
in genomes at high resolution. However, array CGH studies of the human genome noting false negative and false positive results
using large insert clones as probes have raised important concerns regarding the suitability of this approach for clinical
diagnostic applications. Here, we adapt the Smith–Waterman dynamic-programming algorithm to provide a sensitive and robust
analytic approach (SW-ARRAY) for detecting copy-number changes in array CGH data. In a blind series of hybridizations to arrays
consisting of the entire tiling path for the terminal 2 Mb of human chromosome 16p, the method identified all monosomies between
267 and 1567 kb with a high degree of statistical significance and accurately located the boundaries of deletions in the range
267–1052 kb. The approach is unique in offering both a nonparametric segmentation procedure and a nonparametric test of significance.
It is scalable and well-suited to high resolution whole genome array CGH studies that use array probes derived from large
insert clones as well as PCR products and oligonucleotides.
Full-text · Article · Feb 2005 · Nucleic Acids Research