Biology subjects

Egli, A.

Publications and source records attributed to Egli, A..

At least 19 recordsLinked to original sources

bioRxiv2026

A synthetic MALDI-TOF MS atlas for microbial classification across hospitals

Matrix-Assisted Laser Desorption/Ionization Time-of-Flight (MALDI-TOF) mass spectrometry (MS) is central to routine clinical microbiology for species identification. Its combination with machine learning is constrained by fragmented public datasets, species imbalance and acquisition-related variability. Synthetic datasets could provide more training data, but they must preserve the spectral information needed to make predictions on real samples. Here, we present an open, synthetic MALDI-TOF MS atlas comprising 4.3 million spectra across 43 clinically relevant bacterial species. To construct the atlas, we harmonised public repositories from three countries and trained a species- and institution-conditioned concept-bottleneck generative variational autoencoder (CBGM-VAE) on 179,086 measured spectra. We assessed the atlas as both a stand-alone training resource and a complement to real data, using bacterial species classification as a functional benchmark. Evaluation comprised 7,269 real spectra from more than 7 countries, 24 institutions and multiple acquisition platforms, covering: temporal, technical, biological and combined shifts. A neural network trained only on synthetic spectra achieved F1 scores ranging from 0.788 to 1.000 across external cohorts. When synthetic spectra were combined with real training data, classification performance improved or matched using real data only. These findings demonstrate that the atlas retains transferable information and can complement measured data. The openly released atlas provides a reusable resource for supervised learning and benchmarking. It can also be explored for pretraining representation-learning and foundation models in MALDI-TOF MS research.

microbiology↗

bioRxiv2026

Machine learning-guided discovery of a conserved plasmid proteomic signature enables MALDI-TOF MS detection of pOXA-48-carrying Enterobacterales

OXA-48 carbapenemases are among the most widespread and important resistance mechanisms in Enterobacterales. Yet detecting carbapenemases by conventional workflows necessitates additional testing, thus delaying optimization of therapy and implementation of infection control measures. Here, we present a machine learning approach that identifies the conserved pOXA-48 plasmid directly from routine MALDI-TOF spectra acquired for species identification. The model detects pOXA-48 carriers with an AUROC of 0.96-0.98 across two independent hospital cohorts and instrument platforms, indicating near-perfect discrimination. Using bottom-up proteomics, plasmid conjugation, and plasmid curing, we link the discriminative MALDI-TOF spectral features to proteins encoded on pOXA-48, with DUF1496 domain-containing protein producing the most discriminative spectral feature. Our approach reframes the resistance prediction task from inferring a resistance phenotype to detecting a conserved plasmid through its expressed proteomic signature and has the potential to enable rapid MALDI-TOF MS-based diagnostics for a wide range of plasmid-based resistance determinants.

microbiology↗

bioRxiv2026

Competition against the intestinal microbiota selects for pathoadaptive traits in ESBL E. coli

The intestinal tract is a reservoir for Extended-Spectrum {beta}-Lactamase (ESBL)-producing Escherichia coli. Asymptomatic gut colonization by these pathobionts represents a major risk for extraintestinal infections. Despite clinical relevance, the genetic basis of gut colonization by ESBL E. coli remains poorly understood. Here, we determined how the microbiota shapes the fitness landscape of diverse ESBL E. coli strains, defining the functional requirements for intestinal colonization. In microbiota-depleted hosts, colonization relies mostly on metabolic functions. In contrast, in mice harbouring a microbiota, pathoadaptive functions associated with adhesion and biofilm formation are dominant determinants of E. coli fitness, together with accessory virulence functions. Consistent with these observations, experimental evolution in mice reveals convergent adaptation of ESBL E. coli to the presence of a complex microbiota through enhanced adhesion. These findings establish the microbiota as a major ecological driver of pathoadaptation in antibiotic-resistant pathobionts.

microbiology↗

bioRxiv2026

Antimicrobial peptide databases and prediction tools: Toward a standard evaluation framework

Antimicrobial resistance (AMR) has a profound impact on animal and human health and is associated with substantial morbidity, mortality and public health costs. There is a clear need to develop novel, effective antibiotic agents, which can overcome the current AMR crisis. Antimicrobial peptides (AMPs) may offer such a solution and have attracted growing attention for their potential to combat AMR. In parallel, the growing availability of peptide sequences in public databases has stimulated the development of numerous machine learning and deep learning tools to predict antimicrobial activity computationally. However, it remains unclear how reliably these tools can be compared, as existing studies often rely on heterogeneous datasets and inconsistent evaluation protocols that may lead to data leakage and inflated performance estimates. This raises a central question: what evaluation criteria and benchmark resources are needed to enable fair, reproducible, and biologically meaningful assessment of AMP prediction tools? We address this question by focusing specifically on antibacterial peptides (ABPs). We first provide an overview of AMP databases relevant to antibacterial activity and compare their content, redundancy, and experimental metadata. We then critically assess existing computational tools for ABP prediction, highlighting key limitations related to dataset construction, affinity to certain sequences, data leakage, and inconsistent performance reporting. Based on these limitations, we propose a reference evaluation framework designed to improve comparability, reproducibility, and practical utility in ABP prediction. Finally, we provide targeted recommendations for AMP databases and future tool development to support more robust progress in the computational discovery of ABPs.

bioinformatics↗

bioRxiv2025

Ciprofloxacin resistance in Klebsiella pneumoniae: phenotype prediction from genotype and global distribution of resistance determinants

BACKGROUNDCiprofloxacin resistant Klebsiella pneumoniae is common or emerging in many geographies, and knowledge of local resistance rates is important for empirical therapy. Whilst there are known K. pneumoniae ciprofloxacin resistance determinants, there is a lack of systematic data on the effect of determinants, alone and in combination, and there are no publicly accessible tools for predicting resistance from whole genome sequence data. METHODSThe KlebNET-GSP AMR Genotype-Phenotype Group aggregated a matched genotype-phenotype dataset of n=12,167 K. pneumoniae species complex (KpSC) isolates from 27 countries between 2001-2021. We developed a rules-based classifier to predict ciprofloxacin resistance by categorizing the number of quinolone resistance determining regions mutations in gyrA and parC, the number of plasmid-mediated quinolone resistance genes, and the presence/absence of aac(6')-Ib-cr (which can acetylate ciprofloxacin). Predictive performance was assessed using the discovery dataset, for which we re-phenotyped discrepant isolates; and validated using externally contributed datasets (n=7,030 KpSC isolates). RESULTSThe rules-based classifier predicted R vs S/I with categorical agreement, sensitivity, and specificity >96%, and major/very major error rates <4%. Performance was similar across diverse KpSC sources (human, animal, other), species, and intra-species lineages. External validation of the classifier yielded overall 93.12% categorical agreement [95% confidence interval (CI), 92.50-93.74%], 8.65% major errors [95% CI, 7.34-9.97%], and 6.20% very major errors [95% CI, 5.51-6.90%]. We implemented the classifier in Kleborate, a command-line tool that is integrated into the Pathogenwatch web platform. Using this to assess the global distribution of ciprofloxacin resistance determinants in KpSC genomes available in Pathogenwatch (n=31,319, from 109 countries between years 2000-2023), we observed a significant positive association between national quinolone consumption rates and predicted ciprofloxacin resistance (R2=0.20, p=0.004). CONCLUSIONS Ciprofloxacin resistance phenotypes can be reasonably predicted from genotypes, which is sufficient for informing surveillance. However, unexplained resistance remains and accuracy is insufficient for clinical applications. We demonstrate the value of aggregating genotype-phenotype data to explore resistance mechanisms and develop predictors, but highlight complexities in combining phenotype data from different assays and standards.

genomics↗

bioRxiv2025

The natural diversity of E. coli transporter-dependent capsules

Serotyping of bacteria using genomic information (in silico serotyping) has increasingly replaced serology. However, the E. coli capsule serotyping system has been largely abandoned since the 1990s, leaving gaps in our knowledge of capsule genetics, diversity, distribution, and epidemiology. To address this, we established a definitive genotype-serotype map for 35 serologically identified and structurally characterized transporter-dependent capsules. We then surveyed >37,000 E. coli genomes, cataloging 85 transporter-dependent capsule types (K-types), including 55 novel ones. We leveraged this catalog to develop an in silico serotyping tool, kTYPr, and applied it to curated sets of >25,000 E. coli genomes and metagenome-assembled genomes spanning diverse environmental and clinical sources. We found novel K-types enriched in under-sampled environments and associated with E. coli disease. This research expands our understanding of E. coli surface structures, supporting efforts for precision targeting with phage therapy or vaccines.

microbiology↗

bioRxiv2025

Early antifungal resistance prediction based on MALDI-TOF mass spectrometry and machine learning

Antimicrobial resistance (AMR) is a significant global health threat. Recent studies have shown that combining MALDI-TOF mass spectrometry with machine learning (ML) algorithms can accelerate AMR determination. However, these efforts have predominantly focused on bacterial pathogens. Given the significant morbidity, mortality and healthcare costs associated with fungal infections and their evolving antifungal resistance to treatment, there is a need for precision medicine approaches that enable early and accurate detection of antifungal resistance. We developed a machine-learning pipeline that integrates MALDI-TOF mass spectrometry data and drug features to predict antifungal resistance and identify spectral biomarkers for antifungal resistance. By leveraging the DRIAMS dataset, we included 658 pathogen spectra linked to 3,046 phenotypic antifungal resistance results. This dataset covered three drug classes and seven yeast species. The model was trained using categorical, phenotypic antifungal susceptibility testing results as ground truth. Our study systematically investigated the influence of various dimensionality reduction methods on MALDI-TOF mass spectra, antifungal encodings, and machine learning models using nested cross-validation to evaluate predictive performance. We identified that applying principal component analysis to MALDI-TOF mass spectra for dimensionality reduction, and training a multi-layer perceptron yielded the highest and most stable performance for the prediction of antifungal resistance. Our method achieved an AUPRC of 0.77 across the 10 highest-performing species-drug pairs and corresponding Matthews correlation coefficient of 0.69. The model demonstrated the best performance for the species-drug combinations of Candida albicans with micafungin, Candida parapsilosis with fluconazole, and Saccharomyces cerevisiae with itraconazole and fluconazole. By comparing established species-based guidelines, susceptibility test results, and machine learning predictions, we estimated that integrating our algorithm into antifungal selection could help avoid prescriptions to likely resistant pathogens in approximately 3 out of 10 patients for whom standard guidelines recommend such treatments.

bioinformatics↗

bioRxiv2025

Multi-omics profiling of cross-resistance between ceftazidime-avibactam and meropenem identifies common and strain-specific mechanisms in Pseudomonas aeruginosa clinical isolates

Pseudomonas aeruginosa is a highly versatile and resilient pathogen that can infect different tissues and rapidly develop resistance to multiple drugs. Ceftazidime-avibactam (CZA) is an antibiotic often used to treat multidrug resistant infections, however, the knowledge on the CZA resistance mechanisms in P. aeruginosa is limited. Here we performed laboratory evolution of eight clinical isolates of P. aeruginosa exposed to either CZA or meropenem (MEM) in sub-inhibitory concentrations, and used multi-omics profiling to investigate emerging resistance mechanisms. The majority of strains exposed to MEM developed high resistance (83%, 20/24 strains from eight clinical isolates), with only 17% (4/24) acquiring cross-resistance to CZA. The rate of resistance evolution to CZA was substantially lower (21%, 5/24), while 38% (9/24) acquired cross-resistance to MEM. Whole-genome sequencing revealed strain heterogeneity and different evolutionary paths, with three genes mutated in three or more strains: dacB in CZA-treated strains and oprD and ftsI in MEM-treated strains. Transcriptomic and proteomic analysis underlined heterogeneous strain response to antibiotic treatment with few commonly regulated genes and proteins. To identify genes potentially associated with antibiotic resistance, we built a machine learning model that could separate CZA- and MEM-resistant from sensitive strains based on gene expression and protein abundances. To test some of the identified associations, we performed CRISPR/Cas9 genome editing that demonstrated that mutations in dacB, ampD, and to a lesser extent in mexR directly affected CZA resistance. Overall, this study provides novel insights into the strain-specific molecular mechanisms regulating CZA resistance in Pseudomonas aeruginosa. ImportancePseudomonas aeruginosa is one of the most difficult-to-treat pathogens in the hospital, which often acquires resistance to multiple antibiotics. Ceftazidime-avibactam (CZA) is an essential antibiotic used to treat multidrug resistant infections, but its resistance mechanisms are not well understood. Here we investigated the evolution of resistance to CZA and meropenem (MEM) in eight clinical bacterial isolates from patients blood, urine, and sputum. While the rate of resistance evolution to MEM was higher than to CZA, MEM-resistant strains rarely acquired cross-resistance towards CZA. To identify changes at the genome, transcriptome and proteome levels during antibiotic exposure, we performed multi-omics profiling of the evolved strains, and confirmed the effect of several genes on antibiotic resistance with genetic engineering. Altogether, our study provides insights into the molecular response of P. aeruginosa to CZA and MEM and informs therapeutic interventions, suggesting that CZA could still be effective for patients infected with MEM-resistant pathogens.

microbiology↗

bioRxiv2024

Blood-borne immune cells carry low biomass DNA remnants of microbes in patients with colorectal cancer or inflammatory bowel disease

The involvement of the human intestinal microbiome in the regulation of immune cell homeostasis, as well as in the pathogenesis of inflammatory bowel disease (IBD) and colorectal cancer (CRC), are well-established1-4. Bacteria interact with immune cells at the sites of intestinal inflammation, but also in the CRC tumor microenvironment1-6. Indeed, bacterial remnants have recently been detected in human intestinal tissue in patients with IBD, at primary tumor sites and in the metastases of patients with CRC, and in whole blood3,7,8. A defective intestinal epithelial barrier is thought to promote bacterial remnant translocation and disease progression6. However, it is unknown, how bacterial remnants translocate from the intestine to sites of metastasis or into the circulation. We hypothesized that bacterial remnants translocate within peripheral blood immune cells into the circulation. Here, we thus explored the composition of the detectable microbiome in peripheral blood mononuclear cells (PBMCs) of patients with CRC or IBD compared to healthy controls. The PBMC microbiome profiles partially align with the tumor- and metastasis-derived or intestinal tissue-derived microbiome signatures obtained from the same patients with CRC or IBD, respectively. Our metagenomics data, supported by 16S-rRNA-FISH-Flow, imaging flow cytometry and species-specific qPCR, reveal the presence of translocated bacterial genetic sequences in the patients with CRC and IBD, likely due to an intestinal barrier defect. Pathway and serum metabolomics analysis connected to the metabolic potential of the PBMC-derived microbiome sequences support the onset of microbial translocation in such patients. Thus, our data suggest that in patients with intestinal barrier leakage, such as those with CRC or IBD, there is the potential for the translocation of bacterial remnants into the circulation via peripheral immune cells.

cancer biology↗

bioRxiv2024

Chlamydia trachomatis genomes from rectal samples: description of new clade comprising ompA-genotype L4 from Argentina

Whole genome analysis has provided us with insights into the evolution of Chlamydia trachomatis and recently into circulating strains which cause lymphogranuloma venereum (LGV). A large LGV outbreak of a new ompA-genotype, L2b, was first reported in Europe in the early 2000s, primarily affecting men who have sex with men (MSM), and then expanded globally. More recent work shows this outbreak diversifying into variants of described ompA-genotypes, with the same L2b genomic backbone. This study extends the investigation of LGV cases to Argentina and Finland. In 2017, an LGV outbreak was described in Argentina characterized by distinct genomic features shown by both ompA-genotyping and MLST analysis. We have obtained whole genome sequences from cultured isolates and clinical samples via SureSelect (Agilent) target enrichment. Based on ompA and phylogenetic analyses, we describe further diversity within the ompA-genotype L2b clade, illustrating the transmission dynamics in both Argentina and Finland. A key finding is that of a novel clade of Argentinian samples, characterised by a proposed new ompA-genotype L4. Additionally, we present the genome sequence of a non-LGV strain associated with anorectal proctitis. These findings contribute to the investigation of LGV evolution, particularly with the presence of the novel L4 lineage, and provide insights into genomic diversity and transmission dynamics of C. trachomatis. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=99 SRC="FIGDIR/small/624510v1_ufig1.gif" ALT="Figure 1"> View larger version (30K): org.highwire.dtl.DTLVardef@3bd641org.highwire.dtl.DTLVardef@6f6866org.highwire.dtl.DTLVardef@2f1b6eorg.highwire.dtl.DTLVardef@b2137a_HPS_FORMAT_FIGEXP M_FIG C_FIG Impact statementThe phylogeny of C. trachomatis has been described in several publications and has been clear for a decade. The evolution of this STI in the intervening years also informs a great deal about the transmission opportunities and selective pressures that the intracellular bacterium is under. Through genome sequencing, also directly from clinical samples, we provide the first LGV genomes from Argentina and Finland, and describe further evidence of global circulation of the ompA-genotype L2b lineage. We describe a fully novel lineage of LGV C. trachomatis, from Argentina, proposed as ompA-genotype L4. We also find evidence of a strain causing proctitis from the urogenital lineage. Together, these provide significant new findings in the investigation of C. trachomatis. Data summaryAll illumina sequence data, with human read data removed using Hostile (1) and KrakenTools (https://github.com/jenniferlu717/KrakenTools), is deposited with the European Nucleotide Archive (ENA) under project number PRJEB72167.

microbiology↗

bioRxiv2024

The challenge of sequencing Chlamydia trachomatis and other bacterial STI genomes directly from clinical swabs: the optimum solution

Rates of bacterial sexually transmitted infections (STIs) are rising and accessing their genomes provides information on strain evolution, circulating strains, and encoded antimicrobial resistance (AMR). Notable pathogens include Chlamydia trachomatis (CT), Neisseria gonorrhoeae (NG) and Treponema pallidum (TP), globally the most common bacterial STIs. Mycoplasma genitalium (MG) is also a bacterial STI which is of concern due to AMR development. These bacteria are also fastidious or hard to culture, and standard sampling methods lyse bacteria, completely preventing pathogen culture. Clinical samples contain large amounts of human and other microbiota DNA. These factors hinder the sequencing of bacterial STI genomes. We aimed to overcome these challenges in obtaining whole genome sequences, and evaluated four approaches using clinical samples from Argentina (39), Switzerland (14), and cultured samples from Finland (2) and Argentina (1). First, direct genome sequencing from swab samples was attempted through Illumina deep metagenomic sequencing, showing extremely low levels of target DNA, with under 0.01% of the sequenced reads being from the target pathogens. Second, host DNA depletion followed by Illumina sequencing was not found to produce enrichment in these very low load samples. Third, we tried a selective long-read approach with the new adaptive sequencing from Oxford Nanopore Technologies (ONT), which also did not improve enrichment sufficiently to provide genomic information. Finally, target enrichment using a novel pan-genome set of custom SureSelect probes targeting CT, NG, TP, and MG followed by Illumina sequencing was successful. We produced whole genomes from 64% of CT positive samples; from 36% of NG positive samples, and from 60% of TP positive samples. Additionally, we enriched MG DNA to gain partial genomes from 60% of samples. This is the first publication to date to utilize a pan-genome STI panel in target enrichment. Target enrichment, though costly, proved essential for obtaining genomic data from clinical samples. This data can be utilized to examine circulating strains, genotypic resistance, and guide public health strategies. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=126 SRC="FIGDIR/small/624631v1_ufig1.gif" ALT="Figure 1"> View larger version (36K): org.highwire.dtl.DTLVardef@578aeborg.highwire.dtl.DTLVardef@16185f9org.highwire.dtl.DTLVardef@1a2ae87org.highwire.dtl.DTLVardef@1703fa7_HPS_FORMAT_FIGEXP M_FIG C_FIG Impact statementGenome data on circulating sexually transmitted infections (STIs) is important to better understand transmission networks, antimicrobial resistance and to guide treatment decisions. For many bacterial STIs, this information is difficult to obtain, as the bacteria are fastidious, in some cases intracellular, and often recalcitrant to culture. We have developed and tested a target enrichment STI panel of baits to capture whole genomes of Chlamydia trachomatis, Neisseria gonorrhoeae, Treponema pallidum, and Mycoplasma genitalium with approximately 50% success in genome sequencing for the first three pathogens. We compare this against other sequencing and enrichment methods, which did not provide sufficient data for genome analysis. This panel approach shows potential for clinical samples carrying these pathogens and can potentially also be developed for further pathogen groups. Data summaryAll illumina sequence data, with human read data removed using Hostile (1) and KrakenTools (https://github.com/jenniferlu717/KrakenTools), is deposited with the European Nucleotide Archive (ENA) under project number PRJEB72167.

microbiology↗

bioRxiv2024

Multi-layered ecological interactions determine growth of clinical antibiotic-resistant strains within human microbiomes

Understanding the spread of antibiotic-resistant bacteria in human-associated microbial communities, particularly in the gut microbiome, is critical for combating the global resistance crisis. However, the role of various ecological factors, such as intraspecies or interspecies interactions, in modulating this process is poorly understood. We hypothesized that different strains within the same species would exhibit distinct interactions with the resident microbiota, leading to variable invasion outcomes. To test this, we examined the population growth of multiple clinically relevant antibiotic-resistant E. coli strains carrying ESBL and carbapenemase resistance plasmids. We introduced these strains into human gut microbiome samples from healthy individuals, using replicated anaerobic microcosms treated or untreated with antibiotics. We found that while antibiotic exposure significantly influenced the growth of these incoming strains, some were successful even in the absence of antibiotics. The variation in ecological success across strain-microbiome sample combinations was positively associated with the intrinsic growth capacities of the strains in local abiotic conditions and their competitive interactions with resident E. coli. Furthermore, the incoming resistant strains had strain- specific effects on the taxonomic composition of the resident microbiota; different strains not only varied in their growth performance in human microbiomes, in some cases they pushed taxonomic composition of the resident community in different directions. In contrast with phenotypes measured in gut microcosms, metabolic profiles measured across various single-carbon-source environments were not reliable predictors of population growth in human microbiomes. Overall, this controlled design reveals the multi-layered ecological dynamics and strain-specific mechanisms that govern the spread of antibiotic-resistant bacteria in individual human-associated microbiomes.

microbiology↗

bioRxiv2024

Unlocking hidden phage diversity to complete the BASEL collection for systematic exploration of phage-host interactions

Research on bacteriophages, the viruses infecting bacteria, has fueled the development of modern molecular biology and inspired their therapeutic application to combat bacterial multidrug resistance. However, most work has so far focused on a few model phages which impedes direct applications of these findings in clinics and suggests that a vast potential of powerful molecular biology has remained untapped. We have therefore recently composed the BASEL collection of Escherichia coli phages (BActeriophage SElection for your Laboratory) which made a relevant diversity of phages infecting the E. coli K-12 laboratory strain accessible to the community. These phages are widely used, but their assorted diversity had remained limited by the E. coli K-12 host. We have therefore now genetically overcome the two major limitations of E. coli K-12, its lack of O-antigen glycans and the presence of resident bacterial immunity. Restoring O-antigen expression resulted in the isolation of diverse additional viral groups like Kagunavirus, Nonanavirus, Gordonclarkvirinae, and Gamaleyavirus, while eliminating all known antiviral defenses of E. coli K-12 additionally enabled us to isolate phages of Wifcevirus genus. Even though some of these viral groups appear to be common in nature, no phages from any of them had previously been isolated using E. coli laboratory strains and they had thus remained largely understudied. Overall, 37 new phage isolates have been added to complete the BASEL collection. These phages were deeply characterized genomically and phenotypically with regard to host receptors, sensitivity to antiviral defense systems, and host range. These results highlighted dominant roles of the O-antigen barrier for viral host recognition and of restriction-modification systems in bacterial immunity. We anticipate that the completed BASEL collection will propel research on phage-host interactions and their molecular mechanisms, deepening our understanding of viral ecology and fostering innovations in biotechnology and antimicrobial therapy.

microbiology↗

bioRxiv2024

GPT-4 based AI agents - the new expert system for detection of antimicrobial resistance mechanisms?

BackgroundEUCAST recommends a two-step process for beta-lactamases in Gram-negative bacteria. Screening with minimal inhibitory concentrations (MICs) or inhibition zone diameters for potential extended spectrum beta-lactamase (ESBL), plasmid-mediated AmpC beta-lactamase, or carbapenemase production is followed by confirmatory tests. GPT-4 and its newly released customized GPT-agent may support the initial EUCAST-screening process. We aimed to validate a customized GPT-agent to identify potential resistance mechanisms. MethodsWe used 225 Gram-negative isolates. Based on phenotypic resistances against beta-lactam antibiotics, we formed four categories: "none", "ESBL", "AmpC", or "carbapenemase". We included 862 phenotypic categories. Next, we customized a GPT-agent with EUCAST-guidelines, expert rules, and EUCAST-breakpoint table (v13.1). We compared routine diagnostic outputs (reference) to (i) EUCAST-GPT-expert, (ii) medical microbiologists, and (iii) GPT-4 without customization. We determined performance as sensitivities and specificities to flag suspect resistance mechanisms. ResultsThree human readers showed concordance in 814/862 (94.4%) phenotypic categories and used in median eight words (IQR 4-11) for reasoning. Median sensitivity and specificity for ESBL, AmpC, and carbapenemase were 98%/99.1%, 96.8%/97.1%, and 95.5%/98.5%, respectively. Three independent prompting rounds of the GPT-agent showed concordance in 706/862 (81.9%) categories but used in median 158 words (IQR 140-174) for reasoning. Median sensitivity and specificity for ESBL, AmpC, and carbapenemase prediction were 95.4%/69.23%, 96.9%/86.3%, and 100%/98.8%, respectively. In the non-customized GPT-4, 169/862 (19.6%) categories could be interpreted. Of these 137/169 (81.1%) categories agreed with routine diagnostic. The non-customized GPT-4 used in median 85 words (IQR 72-105) for reasoning. ConclusionHuman experts showed higher concordance and shorter argumentations compared to GPT-agents. Human experts showed comparable median sensitivities and higher specificities compared to GPT-agents. GPT-agents showed more unspecific flagging of ESBL and AmpC, potentially, resulting in additional testing, diagnostic delays, and higher costs. GPT-4 and GPT-agents are not IVDR/FDA-approved, but validation of LLMs is critical and datasets for benchmarking are needed.

microbiology↗

bioRxiv2024

A novel barcoded nanopore sequencing workflow of high-quality, full-length bacterial 16S amplicons for taxonomic annotation of bacterial isolates and complex microbial communities

IntroductionDue to recent improvements, Nanopore sequencing has become a promising method for experiments relying on amplicon sequencing. We describe a flexible workflow to generate and annotate high-quality, full-length 16S rDNA amplicons. We evaluated it for two applications, namely, i) identification of bacterial isolates and ii) species-level profiling of microbial communities. MethodsBacterial isolate identification by sequencing was tested on 47 isolates and compared to MALDI-TOF MS. 97 isolates were additionally sequenced to assess the resolution of phylogenetic classification. Species-level community profiling was tested with two full-length 16S primer pairs (A and B) with custom barcodes and compared to results obtained with Illumina sequencing using 27 stool samples. Finally, a Nextflow pipeline was developed to produce high-quality reads and taxonomically annotate them. ResultsWe found high agreement between our workflow and MALDI-TOF data for isolate identification (PPV = 0.90, Cramers V = 0.857 and, Theils U = 0.316). For species-level community profiling, we found strong correlations (rs > 0.6) of alpha diversity indices between the two primer sets and Illumina sequencing. At the community level, we found significant but small differences when comparing sequencing techniques. Finally, we found moderate to strong correlation when comparing relative abundances of individual species (average rs = 0.6 and 0.533, for primers A and B). DiscussionThe proposed workflow enabled accurate identification of single bacterial isolates, making it a worthwhile alternative to MALDI-TOF. While shortcomings have been identified, it enabled reliable identification of prominent features in microbial communities at a fraction of the cost of Illumina sequencing. ImportanceA quick, robust, simple, and cost-effective method to identify bacterial isolates and communities in each sample is indispensable in the fields of microbiology and infection biology. Recent technological advances in Oxford Nanopore Technologies sequencing make this technique an attractive option considering the adaptability, portability, and cost-effectiveness of the platform. Here, we validated a flexible workflow to identify bacterial isolates and characterize bacterial communities using the Oxford Nanopore Technologies sequencing platform combined with the most recent v14 chemistry kits. For bacterial isolates, we compared our nanopore-based approach to MALDI-TOF MS-based identification. For species-level profiling of complex bacterial communities we compared our nanopore-based approach to Illumina shotgun sequencing. For reproducibility purposes, we wrapped the code used to process the sequencing data into a ready-to-use and self-contained Nextflow pipeline.

microbiology↗

bioRxiv2024

Diversity, functional classification and genotyping of SHV β-lactamases in Klebsiella pneumoniae

Interpreting phenotypes of blaSHV alleles in Klebsiella pneumoniae genomes is complex. While all strains are expected to carry a chromosomal copy conferring resistance to ampicillin, they may also carry mutations in chromosomal blaSHV alleles or additional plasmid-borne blaSHV alleles that have extended-spectrum {beta}-lactamase (ESBL) activity and/or {beta}-lactamase inhibitor (BLI) resistance activity. In addition, the role of individual mutations/amino acid changes is not completely documented or understood. This has led to confusion in the literature and in antimicrobial resistance (AMR) gene databases (e.g., NCBIs Reference Gene Catalog and the {beta}-lactamase database (BLDB)) over the specific functionality of individual SHV protein variants. Therefore, identification of ESBL-producing strains from K. pneumoniae genome data is complicated. Here, we reviewed the experimental evidence for the expansion of SHV enzyme function associated with specific amino-acid substitutions. We then systematically assigned SHV alleles to functional classes (wildtype, ESBL, BLI-resistant) based on the presence of these mutations. This resulted in the re-classification of 37 SHV alleles compared with current assignments in NCBIs Reference Gene Catalog and/or BLDB (21 to wildtype, 12 to ESBL, 4 to BLI-resistant). Phylogenetic and comparative genomic analyses support that; i) SHV-1 (encoded by blaSHV-1) is the ancestral chromosomal variant; ii) ESBL and BLI-resistant variants have evolved multiple times through parallel substitution mutations; iii) ESBL variants are mostly mobilised to plasmids; iv) BLI-resistant variants mostly result from mutations in chromosomal blaSHV. We used matched genome-phenotype data from the KlebNET-GSP Genotype-Phenotype Group to identify 3,999 K. pneumoniae isolates carrying one or more blaSHV alleles but no other acquired {beta}-lactamases, with which we assessed genotype-phenotype relationships for blaSHV. This collection includes human, animal, and environmental isolates collected between 2001 to 2021 from 24 countries across six continents. Our analysis supports that mutations at Ambler sites 238 and 179 confer ESBL activity, while most omega-loop substitutions do not. Our data also provide direct support for wildtype assignment of 67 protein variants, including eight that were noted in public databases as ESBL. We reclassified these eight variants as wildtype, because they lack ESBL-associated mutations, and our phenotype data support susceptibility to 3GCs (SHV-27, SHV-38, SHV-40, SHV-41, SHV-42, SHV-65, SHV-164, SHV-187). The approach and results outlined here have been implemented in Kleborate v2.4.1 (a software tool for genotyping K. pneumoniae from genome assemblies), whereby known and novel blaSHV alleles are classified based on causative mutations. Kleborate v2.4.1 was also updated to include ten novel protein variants from the KlebNET-GSP dataset and all alleles in public databases as of November 2023. This study demonstrates the power of sharing AMR phenotypes alongside genome data to improve understanding of resistance mechanisms. Impact statementSince every K. pneumoniae genome has an intrinsic SHV {beta}-lactamase and may also carry additional mobile forms, the correct interpretation of blaSHV genes detected in genome data can be challenging and can lead to K. pneumoniae being misclassified as ESBL-producing. Here, we use matched K. pneumoniae genome and drug susceptibility data contributed from dozens of studies, together with systematic literature review of experimental evidence, to improve our understanding of blaSHV allele variation and mapping of genotype to phenotype. This study shows the value of coordinated data sharing, in this case via the KlebNET-GSP Genotype-Phenotype Group, to improve our understanding of the evolutionary history and functionality of blaSHV genes. The results are captured in an open-source AMR dictionary utilised by the Kleborate genotyping tool, that could easily be incorporated into or used to update other tools and AMR gene databases. This work is part of the wider efforts of the KlebNET-GSP group to develop and support a unified platform tailored for the analysis and interpretation of K. pneumoniae genomes by a wide range of stakeholders. Data summaryBlaSHV allele sequences and class assignments are distributed with Kleborate, v2.4.1, DOI:10.5281/zenodo.10469001. Table S1 provides a summary of blaSHV alleles, including primary accessions, class-modifying mutations, and supporting evidence for class assignments that differ from NCBIs Reference Gene Catalog or BLDB. Whole genome sequence data are publicly available as reads and/or assemblies, individual accessions are given in Table S2; corresponding genotypes and antibiotic susceptibility phenotypes and measurements are available in Tables S3 and S4, respectively.

genomics↗

bioRxiv2023

Type IV-A3 CRISPR-Cas systems drive inter-plasmid conflicts by acquiring spacers in trans

Type IV-A CRISPR-Cas systems are primarily encoded on plasmids and form multi-subunit ribonucleoprotein complexes with unknown biological functions. In contrast to other CRISPR-Cas types, they lack the archetypical CRISPR acquisition module and encode a DinG helicase instead of a nuclease component. Type IV-A3 systems are carried by large conjugative plasmids that often harbor multiple antibiotic-resistance genes. Although their CRISPR array contents suggest a role in inter-plasmid conflicts, this function and the underlying mechanisms have remained unexplored. Here, we demonstrate that a plasmid-encoded type IV-A3 CRISPR-Cas system co-opts the type I-E adaptation machinery from its clinical Klebsiella pneumoniae host to update its CRISPR array. Furthermore, we demonstrate that robust interference of conjugative plasmids and phages is elicited through CRISPR RNA-dependent transcriptional repression. By targeting plasmid core functions, type IV-A3 can prevent the uptake of incoming plasmids, limit their horizontal transfer, and destabilize co-residing plasmids, altogether supporting type IV-A3s involvement in plasmid competition. Collectively, our findings shed light on the molecular mechanisms and ecological function of type IV-A3 systems and have broad implications for understanding and countering the spread of antibiotic resistance in clinically relevant strains.

microbiology↗

bioRxiv2023

Ancient methicillin-resistant Staphylococcus aureus: expanding current knowledge using molecular epidemiological characterization of a Swiss legacy collection

Few methicillin-resistant Staphylococcus aureus (MRSA) from the early years of its global emergence have been sequenced. Knowledge about evolutionary factors promoting the success of specific MRSA multi-locus sequence types (MLSTs) remains scarce. We aimed to characterize a legacy MRSA collection isolated from 1965 to 1987 and compare it against publicly available international and local genomes. We accessed 451 ancient (1965-1987) Swiss MRSA isolates, stored in the Culture Collection of Switzerland. We determined phenotypic antimicrobial resistance (AMR) and performed Illumina short-read sequencing on all isolates and long-read sequencing on a selection with Oxford Nanopore Technology. For context, we included 103 publicly available international genomes from 1960 to 1992 and sequenced 1207 modern Swiss MRSA isolates from 2007 to 2022. We analyzed the core genome (cg)MLST and predicted SCCmec cassette types, AMR, and virulence genes. Among the 451 ancient Swiss MRSA isolates, we found 17 sequence types (STs) of which 11 have been previously described. Two STs were novel combinations of known loci and six isolates carried previously unsubmitted MLST alleles, representing five new STs (ST7843, ST7844, ST7837, ST7839, and ST7842). Most isolates (83% 376/451) represented ST247-MRSA-I isolated in the 1960s, followed by ST7844 (6% 25/451), a novel single locus variant (SLV) of ST239. Analysis by cgMLST indicated that isolates belonging to ST7844-MRSA-III cluster within the diversity of ST239-MRSA-IIII. Early MRSA were predominantly from clonal complex (CC) 8. From 1980 to the end of the 20th century we observed that CC22 and CC5 as well as CC8 were present, both locally and internationally. The combined analysis of 1761 ancient and contemporary MRSA isolates across more than 50 years uncovered novel STs and allowed us a glimpse into the lineage flux between Swiss and international MRSA across time.

microbiology↗