Comparison of library preparation and sequencing depths for direct sequencing of Bordetella pertussis positive samples

Abstract

Whooping cough, or pertussis, is a highly transmissible respiratory infection caused by Bordetella pertussis. Due to the high burden of pertussis, vaccine programmes were introduced internationally and in Australia since the 1950s. This has resulted in a significant decrease of pertussis infections. However, since the 1990s the number of pertussis notifications has increased considerably. Currently circulating B. pertussis strains differ in vaccine antigen composition compared to strains that circulated in the pre-vaccination era. These genetic differences are thought to contribute, in part, to the re-emergence of pertussis in Australia and around the world. Whole genome sequencing (WGS) can resolve minute differences in circulating strains and provides unparalleled resolution of vaccine antigens. This high-resolution snapshot can provide clues that enable more targeted public health interventions. However, pertussis is primarily diagnosed with culture-independent diagnostic assays which offer fast turnaround result times and reduced laboratory costs, eliminating the need to culture isolates. Current WGS methods require a cultured isolate, resulting in an absence of B. pertussis genome sequences in the post vaccination era. This scarcity has, in turn, limited understanding of currently circulating strains and respective vaccine antigen compositions. Recent advancements of WGS technologies have allowed direct sequencing of clinical specimens without the need for a cultured isolate. However, recovering reliable sequence data from clinical samples of low bacterial load infections such as B. pertussis is a pressing challenge. We sought to increase the yield of B. pertussis sequences direct from a clinical sample by evaluating widely available WGS library preparation methods. We report that the Illumina DNA prep library preparation kit combined with deep sequencing allowed the detection of important surveillance information such as allelic variations in the B. pertussis vaccine antigens. Further, our method generates high coverage over the 23S ribosomal RNA of B. pertussis enabling macrolide resistance to be easily determined. Overall, this method can improve surveillance of B. pertussis, by monitoring changes in vaccine antigens, detecting antimicrobial resistance and guiding Public Health control interventions.