Identification of intermediate-sized deletions and inference of their impact on gene expression in a human population

• Main Authors: Jing Hao Wong, Daichi Shigemizu, Yukiko Yoshii, Shintaro Akiyama, Azusa Tanaka, Hidewaki Nakagawa, Shu Narumiya, Akihiro Fujimoto
• Format: Article
• Language: English
• Published: BMC 2019-07-01
• Series: Genome Medicine
• Subjects: Intermediate-sized deletion; Expression quantitative trait loci (eQTL); Genomic imputation; Long-read sequencing
• Online Access: http://link.springer.com/article/10.1186/s13073-019-0656-4

Abstract Background Next-generation sequencing has allowed for the identification of different genetic variations, which are known to contribute to diseases. Of these, insertions and deletions are the second most abundant type of variations in the genome, but their biological importance or disease association is not well-studied, especially for deletions of intermediate sizes. Methods We identified intermediate-sized deletions from whole-genome sequencing (WGS) data of Japanese samples (n = 174) with a novel deletion calling method which considered multiple samples. These deletions were used to construct a reference panel for use in imputation. Imputation was then conducted using the reference panel and data from 82 publically available Japanese samples with gene expression data. The accuracy of the deletion calling and imputation was examined with Nanopore long-read sequencing technology. We also conducted an expression quantitative trait loci (eQTL) association analysis using the deletions to infer their functional impacts on genes, before characterizing the deletions causal for gene expression level changes. Results We obtained a set of polymorphic 4378 high-confidence deletions and constructed a reference panel. The deletions were successfully imputed into the Japanese samples with high accuracy (97.3%). The eQTL analysis identified 181 deletions (4.1%) suggested as causal for gene expression level changes. The causal deletion candidates were significantly enriched in promoters, super-enhancers, and transcription elongation chromatin states. Generation of deletions in a cell line with the CRISPR-Cas9 system confirmed that they were indeed causative variants for gene expression change. Furthermore, one of the deletions was observed to affect the gene expression levels of a gene it was not located in. Conclusions This paper reports an accurate deletion calling method for genotype imputation at the whole genome level and shows the importance of intermediate-sized deletions in the human population.