PDFs of the slides are available at: https://github.com/klebgenomics/KlebNetTrainingSep2025

A version of these slides were presented live by KlebNET team members as part of the GenEpi-BioTrain – Genetic Epidemiology and Bioinformatics Training Programme run by the European CDC (ECDC).

To see/hear the video recordings of the KlebNET lectures you need to enrol (free!) in the GenEpi-BioTrain and navigate to the Klebsiella session.

Lecture topics

1. Introduction to Klebsiella pneumoniae – Kat Holt, London School of Hygiene and Tropical Medicine, UK

2. Bacterial strain taxonomy using LIN codes – Sylvain Brisse, Institut Pasteur, Paris, France

3. In silico serotyping of Klebsiella – Tom Stanton, Monash University, Australia

4. K. pneumoniae genome analysis with Kleborate – Margaret Lam, Monash University, Australia

5. Convergence of hypervirulence and AMR – Margaret Lam, Monash University, Australia

6. Seroepidemiology of Klebsiella – Kelly Wyres, Monash University, Australia

7. One Health genomic surveillance of K. pneumoniae – Marit Hetland, Stavanger University Hospital, Norway

High-throughput massive parallel sequencing has significantly improved bacterial pathogen genomics, diagnostics, and epidemiology. Despite its high accuracy, short-read sequencing struggles with complete genome reconstruction and assembly of extrachromosomal elements such as plasmids. Long-read sequencing with Oxford Nanopore Technologies (ONT) presents an alternative that offers benefits including real-time sequencing and cost-efficiency, particularly useful in resource-limited settings. However, the historically higher error rates of ONT data have so far limited its application in high-precision genomic typing. The recent release of ONT’s R10.4.1 chemistry, with significantly improved raw read accuracy (Q20+), offers a potential solution to this problem. The aim of this study was to evaluate the performance of ONT’s latest chemistry for bacterial genomic typing against the gold standard Illumina technology, focusing on three respiratory pathogens of public health importance, Klebsiella pneumoniae, Bordetella pertussis, and Corynebacterium diphtheriae, and their related species. Using the Rapid Barcoding Kit V14, we generated and analyzed genome assemblies with different basecalling models, at different simulated depths of coverage. ONT assemblies were compared to the Illumina reference for completeness and core genome multilocus sequence typing (cgMLST) accuracy (number of allelic mismatches). Our results show that genomes obtained from raw ONT data basecalled with Dorado SUP v0.9.0, assembled with Flye, and with a minimum coverage depth of 35×, optimized accuracy for all bacterial species tested. Error rates were consistently below 0.5% for each cgMLST scheme, indicating that ONT R10.4.1 data is suitable for high-resolution genomic typing applied to outbreak investigations and public health surveillance.

Read the paper here: https://pubmed.ncbi.nlm.nih.gov/40456603/

The functionality is geared towards exploring sets of K/O antigens, in terms of their prevalence and distribution across geographical regions and theoretical coverage of infection isolates, to inform vaccine design.

Prevalence estimates are adjusted for localised nosocomial clustering, to reduce the bias introduced by random outbreaks during surveillance periods. Coverage estimates are based on total isolates, not adjusted for clustering.

The Modelled antigen prevalence tab is populated with pre-calculated global and regional prevalence estimates modelled using Bayesian hierarchical meta-analysis, as described in the paper. Subgroup analyses are limited to those modelled and reported in the paper (geographic regions, fatal cases, ESBL- or carbapenemase- carrying isolates).

The Dynamic antigen prevalence tab is populated with simple pooled estimates calculated on the fly, allowing users to interactively explore prevalence and coverage more flexibly by country, study, year, and multi-locus sequence type (ST).

Full Resource: https://klebsiella.shinyapps.io/neonatal/

Background Klebsiella pneumoniae causes ∼20% of sepsis in neonates, with ∼40% crude mortality. A vaccine administered to pregnant women, protecting against ≥70% of K. pneumoniae infections, could avert ∼400,000 cases and ∼80,000 deaths annually, mostly in Africa and South Asia. Vaccine formulations targeting the capsular polysaccharide (K) or lipopolysaccharide (O) antigens are in development. Global K. pneumoniae populations display extensive K and O diversity, necessitating a polyvalent vaccine targeted to the serotypes associated with neonatal disease in relevant geographical regions. We investigated the prevalence of K and O types associated with neonatal sepsis in Africa and South Asia to inform maternal vaccine design.

Methods and Findings We analysed n=1930 K. pneumoniae neonate blood isolates from 13 surveillance studies across 35 sites in 13 countries. We used pathogen sequencing to predict K and O serotypes and correct for local transmission clusters, and Bayesian hierarchical meta-analysis to estimate K and O prevalence. Eighty-seven K loci were identified. KL2, KL102, KL25, KL15 and KL62 accounted for 49% of isolates. We estimate that 20 K loci, combining the eight most prevalent per region, could cover 72.9% of all infections [95% credible interval, 69.4–76.5%] and ≥70% in each of Eastern, Western and Southern Africa and South Asia. Preliminary findings from three sites suggested sufficient temporal stability of K loci to maintain 20-valent K vaccine coverage over 5-10 years, but more longitudinal data are needed to support this prediction. O types were far less diverse (n=14 types). We estimate the top-5 (O1⍺β,2⍺, O1⍺β,2β, O2⍺, O2β and O4) would cover 86% [82.6–89.9%] of total infections (76–92% per region), while the top-10 would cover ∼99% of infections in all four regions.

Conclusions Neonatal sepsis is associated with diverse K and O types, with substantial geographic and temporal variation even after adjusting for localised transmission clusters. Despite this, a single 20-valent K vaccine could theoretically cover ≥70% of infections in all target regions. Locally-targeted vaccines could achieve higher coverage with lower valency, but are less feasible. In principle, very high coverage could be achieved with lower valency O-based vaccines, however protective efficacy of antibodies targeting the O antigen remains uncertain. Further research is needed on cross-reactivity, antigen exposure and stability of antigens over time, to better inform vaccine development.

Full Printout

https://www.medrxiv.org/content/10.1101/2025.06.28.25330253v1.full.pdf

Klebsiella pneumoniae is a priority pathogen for the development of novel therapeutic and prophylactic measures such as vaccines and monoclonal antibody therapies. Key molecular targets include the polysaccharide capsule, which protects K. pneumoniae from phagocytosis and serum bactericidal activity, and is highly immunogenic. A total of 77 distinct serological phenotypes have been defined but more than 160 different capsule polysaccharides are predicted based on analysis of the corresponding capsule (K) biosynthesis loci from whole genome sequences (WGS). These loci form the basis for WGS-based capsule typing with the computational tool, Kaptive, enabling capsule epidemiology investigations at scale to inform vaccine design. However, to-date limited data has been available with which to assess the accuracy of Kaptive for phenotype inference. 

Here we present a preliminary analysis of 731 matched serological and genomic K-type calls, and show that 84.5% of the serotypes are concordant with genomic predictions, supporting use of Kaptive for seroepidemiology analyses. Ongoing work will expand the dataset, investigate genotype-phenotype discrepancies and update Kaptive to further improve its accuracy.

Full Technical Report：https://zenodo.org/records/15742130

Surface polysaccharides are common antigens in priority pathogens and therefore attractive targets for novel control strategies such as vaccines, monoclonal antibody and phage therapies. Distinct serotypes correspond to diverse polysaccharide structures that are encoded by distinct biosynthesis gene clusters; e.g. the Klebsiella pneumoniae species complex (KpSC) K- and O-loci encode the synthesis machinery for the capsule (K) and outer-lipopolysaccharides (O), respectively. We previously presented Kaptive and Kaptive 2, programmes to identify K- and O-loci directly from KpSC genome assemblies (later adapted for Acinetobacter baumannii), enabling sero-epidemiological analyses to guide vaccine and phage therapy development. However, for some KpSC genome collections, Kaptive (v≤2) was unable to type a high proportion of K-loci. Here, we identify the cause of this issue as assembly fragmentation and present a new version of Kaptive (v3) to circumvent this problem, reduce processing times and simplify output interpretation. We compared the performance of Kaptive v2 and Kaptive v3 for typing genome assemblies generated from subsampled Illumina read sets (decrements of 10× depth), for which a corresponding high-quality completed genome was also available to determine the ‘true’ loci (n=549 KpSC, n=198 A. baumannii). Both versions of Kaptive showed high rates of agreement to the matched true locus amongst ‘typeable’ locus calls (≥96% for ≥20× read depth), but Kaptive v3 was more sensitive, particularly for low-depth assemblies (at <40× depth, v3 ranged 0.85–1 vs v2 0.09–0.94) and/or typing KpSC K-loci (e.g. 0.97 vs 0.82 for non-subsampled assemblies). Overall, Kaptive v3 was also associated with a higher rate of optimal outcomes; i.e. loci matching those in the reference database were correctly typed, and genuine novel loci were reported as untypeable (73–98% for v3 vs 7–77% for v2 for KpSC K-loci). Kaptive v3 was >1 order of magnitude faster than Kaptive v2, making it easy to analyse thousands of assemblies on a desktop computer, facilitating broadly accessible in silico serotyping that is both accurate and sensitive. The Kaptive v3 source code is freely available on GitHub (https://github.com/klebgenomics/Kaptive), and has been implemented in Kaptive Web (https://kaptive-web.erc.monash.edu/).

Read the paper here: Fast and accurate in silico antigen typing with Kaptive 3

Version 3 is faster, mainly because of updates to Kaptive which we use for capsule (K) and O typing. We have refactored the code base to introduce a modular approach, making Kleborate more flexible to run and easier to develop. Look out for new modules for Klebsiella oxytoca and E. coli – coming soon.

Kleborate v3 is available in GitHub and is included in the latest versions of Pathogenwatch and Bactopia, and the Kleborate tutorial.

Thanks to Mary Maranga for doing the bulk of the coding, and also to Ryan Wick for kicking on the modular rewrite, and Kara Tsang & Maggie Lam for updating and testing AMR & virulence databases.