Standardization and Diversification of Copy Number Microarray Testing for Clinical Diagnostics—Implications of the Cross-Platform/Algorithm Study on Clinical Diagnostic Chromosomal Microarray Analysis

Abstract

Chromosomal microarray analysis (CMA)6 has rapidly transitioned from the research setting to the clinical diagnostic setting, owing to its high sensitivity, specificity, and reliability, and by proving to be a substantial improvement over conventional cytogenetic analysis (e.g., karyotyping). Within 3 years, CMA has moved from replacing the “stepwise approach of multiple FISH studies in patients with unclear etiologies” and “an adjunct to conventional cytogenetics” (1) to “the first tier test for patients with developmental delay/intellectual disability, multiple congenital anomalies and/or autism spectrum disorders” (2). CMA has had a high impact in clinical diagnostics, leading to the discovery of new genomic disorders, and has become an indispensable tool for routine molecular and cytogenetic testing. Nonetheless, our understanding of all the nuances that influence the performance characteristics of CMA is incomplete, as was illustrated in a recent report in Nature Biotechnology that assessed performance across platforms and algorithms (3). G-banding karyotyping, the dominant form of conventional cytogenetic analysis, uniformly interrogates the entire genome at once. In contrast, CMA involves the use of several different types of microarrays and interrogates the entire genome by selectively sampling genomic loci with specific probes. Differences in probe type, probe location, and probe density (interprobe spacing) exist among array platforms. The Nature Biotechnology report provides the most comprehensive cross-platform comparison and uncovers important clinical implications for the applications of CMA, particularly with respect to array design and copy number variation (CNV) detection. There are 2 basic array types—one that originated from the comparative genomic hybridization (CGH) technique and the other from the single-nucleotide polymorphism (SNP) genotyping array. Both array platforms have undergone a series of rapid improvements. For example, Agilent CGH arrays evolved from 44K arrays (44 000 probes on the array) to 105K arrays, and then to 180K, 244K, and 1000K arrays (1 000 000 …