Comparative Analysis for the Performance of Variant Calling Pipelines on Detecting the de novo Mutations in Humans

• Main Authors: Yu Liang, Li He, Yiru Zhao, Yinyi Hao, Yifan Zhou, Menglong Li, Chuan Li, Xuemei Pu, Zhining Wen
• Format: Article
• Language: English
• Published: Frontiers Media S.A. 2019-04-01
• Series: Frontiers in Pharmacology
• Subjects: de novo mutation; rare diseases; variant calling pipelines evaluation; gene function; whole-exon sequencing
• Online Access: https://www.frontiersin.org/article/10.3389/fphar.2019.00358/full

Despite of the low occurrence rate in the entire genomes, de novo mutation is proved to be deleterious and will lead to severe genetic diseases via impacting on the gene function. Considering the fact that the traditional family based linkage approaches and the genome-wide association studies are unsuitable for identifying the de novo mutations, in recent years, several pipelines have been proposed to detect them based on the whole-genome or whole-exome sequencing data and were used for calling them in the rare diseases. However, how the performance of these variant calling pipelines on detecting the de novo mutations is still unexplored. For the purpose of facilitating the appropriate choice of the pipelines and reducing the false positive rate, in this study, we thoroughly evaluated the performance of the commonly used trio calling methods on the detection of the de novo single-nucleotide variants (DNSNVs) by conducting a comparative analysis for the calling results. Our results exhibited that different pipelines have a specific tendency to detect the DNSNVs in the genomic regions with different GC contents. Additionally, to refine the calling results for a single pipeline, our proposed filter achieved satisfied results, indicating that the read coverage at the mutation positions can be used as an effective index to identify the high-confidence DNSNVs. Our findings should be good support for the committees to choose an appropriate way to explore the de novo mutations for the rare diseases.