Quality Assessment and Validation of High-Throughput Sequencing for Grapevine Virus Diagnostics

Abstract

Development of High-Throughput Sequencing (HTS), also known as next generation sequencing, revolutionized diagnostic research of plant viruses. HTS outperforms bioassays and molecular diagnostic assays that are used to screen domestic and quarantine grapevine materials in data throughput, cost, scalability, and detection of novel and highly variant virus species. However, before HTS-based assays can be routinely used for plant virus diagnostics, performance specifications need to be developed and assessed. In this study, we selected 18 virus-infected grapevines as a test panel for measuring performance characteristics of an HTS-based diagnostic assay. Total nucleic acid (TNA) was extracted from petioles and dormant canes of individual samples and constructed libraries were run on Illumina NextSeq 500 instrument using a 75-bp single-end read platform. Sensitivity was 98% measured over 264 distinct virus and viroid infections with a false discovery rate (FDR) of approximately 1 in 5 positives. The results also showed that combining a spring petiole test with a fall cane test increased sensitivity to 100% for this TNA HTS assay. To evaluate extraction methodology, these results were compared to parallel dsRNA extractions. In addition, in a more detailed dilution study, the TNA HTS assay described here consistently performed well down to a dilution of 5%. In that range, sensitivity was 98% with a corresponding FDR of approximately 1 in 5. Repeatability and reproducibility were assessed at 99% and 93%, respectively. The protocol, criteria, and performance levels described here may help to standardize HTS for quality assurance and accreditation purposes in plant quarantine or certification programs.

Keywords: grapevine virus; high-throughput sequencing; performance assessment; repeatability; reproducibility; sensitivity; specificity; validation.

Figure 1

Average Ct values of endornaviruses PvEV-1, PvEV-2 from extracted TNA of common bean leaves over a 40-day developmental period. The bars indicate standard error.

Figure 2

Impact of dilutions on grapevine virus read counts. The average count of reads mapped to grapevine virus sequences is averaged for and plotted against dilution. Regression through the origin demonstrates a high degree of linearity (R² > 0.99).