Detecting copy-number variations in whole-exome sequencing data using the eXome Hidden Markov Model: an ‘exome-first’ approach

Abstract

Whole-exome sequencing (WES) is becoming a standard tool for detecting nucleotide changes, and determining whether WES data can be used for the detection of copy-number variations (CNVs) is of interest. To date, several algorithms have been developed for such analyses, although verification is needed to establish if they fit well for the appropriate purpose, depending on the characteristics of each algorithm. Here, we performed WES CNV analysis using the eXome Hidden Markov Model (XHMM). We validated its performance using 27 rare CNVs previously identified by microarray as positive controls, finding that the detection rate was 59%, or higher (89%) with three or more targets. XHMM can be effectively used, especially for the detection of >200 kb CNVs. XHMM may be useful for deletion breakpoint detection. Next, we applied XHMM to genetically unsolved patients, demonstrating successful identification of pathogenic CNVs: 1.5-1.9-Mb deletions involving NSD1 in patients with unknown overgrowth syndrome leading to the diagnosis of Sotos syndrome, and 6.4-Mb duplication involving MECP2 in affected brothers with late-onset spasm and progressive cerebral/cerebellar atrophy confirming the clinical suspect of MECP2 duplication syndrome. The possibility of an 'exome-first' approach for clinical genetic investigation may be considered to save the cost of multiple investigations.

Expanding Genetic and Functional Diagnoses of <b><i>IGF1R</i></b> Haploinsufficiencies

“…Of these, only one, a 4.492 Mb heterozygous chromosome 15 deletion that spans 28 genes, including the entire IGF1R gene, was unique to P3. Our finding of an IGFIR deletion in P3, therefore, supports the efficacy and efficiency of using exome sequencing for both variant and CNV detection, and is in-line with the proposed “exome-first” paradigm for clinical investigations of genetic defects (32). While successfully applied to our present case, additional algorithm refinement and testing will likely be necessary before these approaches can be widely applied clinically in the genetic workup for patients with rare diseases.…”

“…For P3, we leveraged existing WES data to serve as a basis for CNV discovery, an application that provides much finer resolution (less than 1 kb) than conventional CNV methods. WES data generated from patient P3 and her parents had not identified candidate single nucleotide variations of significance, but through XHMM analysis (17), which is a relatively sensitive and specific algorithm for detecting rare, large, CNVs (29–31), 27 CNVs in P3 were uncovered. Of these, only one, a 4.492 Mb heterozygous chromosome 15 deletion that spans 28 genes, including the entire IGF1R gene, was unique to P3.…”

Novel Genetic Triggers and Genotype–Phenotype Correlations in Patients With Left Ventricular Noncompaction

“…The disease cohort was also significantly enriched for multiple VOIs in individual patients; approximately 1.7 as many patients had at least one VOI in our disease cohort, but 4.2 times as many individuals had 2 or more VOIs (Figure S2B in the Supplementary Appendix). Copy number analysis was performed from the WES data using XHMM 30 and no high confidence copy number variants were identified in known cardiomyopathy genes.…”

Whole-exome sequencing and digital PCR identified a novel compound heterozygous mutation in the NPHP1 gene in a case of Joubert syndrome and related disorders

“…Although several algorithms capable of identifying CNVs from WES data have been proposed, there is no WES-based CNV-calling method that can detect all CNVs in a large size range [13–15]. Because non-allelic homologous recombination between low-copy repeats on chromosome 2q13 have been shown to be a major cause of large deletions involving the NPHP1 gene [11, 16, 17], we chose to use XHMM based CNV-calling algorithm, which is suitable for predicting CNVs ranging from 10 kb to 1 Mb [13].…”