Penncnv-Exomeseq: Genotype Improves Copy Number Variant Detection in Exome Sequencing

Background: Sequencing has become a popular method for the generation of large-scale genomic data and with the inundation of such data source comes the necessity for accurate genotype calling of nucleotide bases (A/T/C/G) and copy number (0/1/2/3/4) variants (CNV). The use of SNP arrays as a point of reference for widely used assays for genomic variants, including the variability of different centers and algorithms impacting quality may bear fruit. PennCNV is the most popular method for CNV detection from SNP arrays. Therefore, we observe the unique features that set it apart: namely using both intensity and genotype in tandem to infer CNV states using an HMM and trio based recalling of CNVs to bring de novo rates to an acceptably low level. Results: Sequencing offers features to assess CNVs intensity which has been leveraged by a number of algorithms, including XHMM, but the valuable feature of genotype for call accuracy has not been incorporated. Here we show derivation of genotype frequency from exome sequencing as a robust data to supplement intensity data in CNV detection.We detect more CNVs at a higher true positive rate than existing methods. Conclusion: This application of BAF furthermore allows an arsenal of tools to be utilized including PennCNV and ParseCNV for sequencing data. PennCNVExomeSeq is freely available