Abstract
BackgroundCopy number variants (CNVs) are known to play an important role in the development and progression of several diseases. However, detection of CNVs with whole-exome sequencing (WES) experiments is challenging. Usually, additional experiments have to be performed.FindingsWe developed a novel algorithm for somatic CNV calling in matched WES data called “CopyDetective". Different from other approaches, CNV calling with CopyDetective consists of a 2-step procedure: first, quality analysis is performed, determining individual detection thresholds for every sample. Second, actual CNV calling on the basis of the previously determined thresholds is performed. Our algorithm evaluates the change in variant allele frequency of polymorphisms and reports the fraction of affected cells for every CNV. Analyzing 4 WES data sets (n = 100) we observed superior performance of CopyDetective compared with ExomeCNV, VarScan2, ControlFREEC, ExomeDepth, and CNV-seq.ConclusionsIndividual detection thresholds reveal that not every WES data set is equally apt for CNV calling. Initial quality analyses, determining individual detection thresholds—as realized by CopyDetective—can and should be performed prior to actual variant calling.
Highlights
Copy number variants (CNVs) are known to play an important role in the development and progression of several diseases
Performance is compared to 5 established tools for CNV calling in whole-exome sequencing (WES) data: ExomeCNV, VarScan2, ExomeDepth, Control-FREEC (WR and KS), and CNV-seq
It should be noted that loss of heterozygosity (LOH) was excluded from Data Set 4 because we do not have any information on the frequency of affected cells for these calls
Summary
Copy number variants (CNVs) are known to play an important role in the development and progression of several diseases. Conclusions: Individual detection thresholds reveal that not every WES data set is apt for CNV calling. Initial quality analyses, determining individual detection thresholds—as realized by CopyDetective—can and should be performed prior to actual variant calling. Despite continuously decreasing costs for experiments, it is desirable to keep the number of necessary genetic experiments to a minimum—not least because of limited tumor material [5,6]. It would be most practical if there were valid algorithms to determine single-nucleotide variants (SNVs), short insertions and deletions (indels), structural variants (SVs), and copy number variants (CNVs) by just a single NGS experiment.
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call