Targeted long-read sequencing allows for rapid identification of pathogenic disease-causing variants in retinoblastoma

Abstract

ABSTRACT Background Identification of disease-causing variants of the retinoblastoma gene (RB1), the predominant cause of retinoblastoma, is challenging. Targeted long-read genome sequencing offers a novel approach to resolve the diverse range of pathogenic variants in RB1 and provides haplotype information rapidly. Materials and Methods Genomic DNA was isolated from a venipuncture blood draw of a retinoblastoma patient. Whole genome sequencing (WGS) was carried out using the short-read Ilumina platform. WGS and targeted sequencing of RB1 was accomplished using the long-read Oxford Nanopore Technologies (ONT) platform. Deep-learning frameworks allowed haplotagging, variant calling, and variant annotation of both short- and long-read data. Results Targeted long-read sequencing of the RB1 gene allowed for enhanced depth of read coverage for discovery of rare variants and haplotype analysis. A duplication leading to a frameshift and early termination in RB1 was identified as the most deleterious variant by all sequencing methods, with long-read technology providing additional information of methylation signal and haplotype information. More importantly, there was greater than 98% concordance of RB1 variants identified between short-read and targeted long-read sequencing modalities. Conclusions Targeted long-read technology allows for focused sequencing effort for variant discovery. Application of this for the first time in a retinoblastoma patient allowed haplotagged variant identification and demonstrated excellent concordance with benchmark short-read sequencing. The added benefit of targeted long-read sequencing to resolve disease-causing genomic variation in RB1 rapidly from a blood draw will provide a more definitive diagnosis of heritable RB and guide management decisions for patients and their families.