Motivation: Only a few large systematic studies have evaluated the impact of copy number variants (CNVs) on common diseases. Several million individuals have been genotyped on single nucleotide variation arrays, which could be used for genome-wide CNVs association studies. However, CNV calls remain prone to false positives and only empirical filtering strategies exist in the literature. To overcome this issue, we defined a new quality score (QS) estimating the probability of a CNV called by PennCNV to be confirmed by other software.
Results: Out-of-sample comparison showed that the correlation between the consensus CNV status and the QS is twice as high as it is for any previously proposed CNV filters. ROC curves displayed an AUC higher than 0.8 and simulations showed an increase up to 20% in statistical power when using QS in comparison to other filtering strategies. Superior performance was confirmed also for alternative consensus CNV definition and through improving known CNV-trait associations.
Availability and implementation: http://goo.gl/T6yuFM CONTACT: firstname.lastname@example.org or aurelien@email@example.comSupplementary information: Supplementary data are available at Bioinformatics online.
© The Author 2016. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: firstname.lastname@example.org.