PCR-free, targeted genomic sequencing using Dynamically optimized reference Adaptive Sampling (DORAS)

Abstract

Whole genome sequencing (WGS) has become a cornerstone of clinical microbiology, enabling comprehensive analysis of microbial genome diversity. However, WGS is often computationally intensive and time-consuming when applied to specific applications like multilocus sequence typing (MLST), where only a subset of genes is only needed for typing. This study evaluates the potential of adaptive sampling (AS), a software-based solution available on Oxford Nanopore Technologies (ONT) devices, to optimize sequencing runs for MLST by reducing the production of unnecessary reads falling outside of the target areas. We demonstrate that AS, when used directly with the target gene sequences, does not reach sufficient target coverage when compared to WGS baseline sequencing due to inefficient read recruitment. Thus, we developed a novel, PCR-free approach, termed Dynamically Optimized Reference Adaptive Sampling (DORAS), which streamlines gene-specific enrichment by targeting genomic regions of interest and their genomic vicinity. DORAS first determines the genomic context of regions of interest for each sample, and then dynamically adjusts the length of the reference sequences during live sequencing. Consensus sequences are periodically constructed and evaluated for taxonomic classification. We demonstrate that full MLST profiles can be obtained in approximately half the time required for whole-genome sequencing to achieve 30X coverage (3 vs. 6 h), with no additional hands-on library preparation time. Validation on clinical isolates from hospital outbreaks belonging to Corynebacterium diphtheriae, vancomycin-resistant Enterococci, and routine clinical E. coli isolates, demonstrated the consistent retrieval of MLST types as compared to standard WGS methods. DORAS thus offers a cost-effective, efficient solution for routine surveillance and outbreak investigations based on MLST types in the clinical setting.