NaHCO3 responsiveness is a novel phenotype where some methicillin-resistant Staphylococcus aureus (MRSA) isolates exhibit significantly lower minimal inhibitory concentrations (MIC) to oxacillin and/or cefazolin in the presence of NaHCO3. NaHCO3 responsiveness correlated with treatment response to β-lactams in an endocarditis animal model. We investigated whether treatment of NaHCO3-responsive strains with β-lactams was associated with faster clearance of bacteremia. The CAMERA2 trial (Combination Antibiotics for Methicillin-Resistant Staphylococcus aureus) randomly assigned participants with MRSA bloodstream infections to standard therapy, or to standard therapy plus an anti-staphylococcal β-lactam (combination therapy). For 117 CAMERA2 MRSA isolates, we determined by broth microdilution the MIC of cefazolin and oxacillin, with and without 44 mM of NaHCO3. Isolates exhibiting ≥4-fold decrease in the MIC to cefazolin or oxacillin in the presence of NaHCO3 were considered \"NaHCO3-responsive\" to that agent. We compared the rate of persistent bacteremia among participants who had infections caused by NaHCO3-responsive and non-responsive strains, and that were assigned to combination treatment with a β-lactam. Thirty-one percent (36/117) and 25% (21/85) of MRSA isolates were NaHCO3-responsive to cefazolin and oxacillin, respectively. The NaHCO3-responsive phenotype was significantly associated with sequence type 93, SCCmec type IVa, and mecA alleles with substitutions in positions -7 and -38 in the regulatory region. Among participants treated with a β-lactam, there was no association between the NaHCO3-responsive phenotype and persistent bacteremia (cefazolin, P = 0.82; oxacillin, P = 0.81). In patients from a randomized clinical trial with MRSA bloodstream infection, isolates with an in vitro β-lactam-NaHCO3-responsive phenotype were associated with distinctive genetic signatures, but not with a shorter duration of bacteremia among those treated with a β-lactam.

This computational protocol describes how to use pyPGCF, a python software package that runs in the linux environment, in order to analyze bacterial genomes and perform: (i) phylogenomic analysis, (ii) species demarcation, (iii) identification of the core proteins of a bacterial genus and its individual species, (iv) identification of species-specific fingerprint proteins that are found in all strains of a species and, at the same time, are absent from all other species of the genus, (v) functional annotation of the core and fingerprint proteins with eggNOG, and (vi) identification of secondary metabolite biosynthetic gene clusters (smBGCs) with antiSMASH. This software has already been implemented to analyze bacterial genera and species that are important for plants (e.g., Pseudomonas, Bacillus, Streptomyces). In addition, we provide a test dataset and example commands showing how to analyze 165 genomes from 55 species of the genus Bacillus. The main advantages of pyPGCF are that: (i) it uses adjustable orthology cut-offs, (ii) it identifies species-specific fingerprints, and (iii) its computational cost scales linearly with the number of genomes being analyzed. Therefore, pyPGCF is able to deal with a very large number of bacterial genomes, in reasonable timescales, using widely available levels of computing power.

Source attribution has traditionally involved combining epidemiological data with different pathogen characterisation methods, including 7-gene multi locus sequence typing (MLST) or serotyping, however, these approaches have limited resolution. In contrast, whole genome sequencing data provide an overview of the whole genome that can be used by attribution algorithms. Here, we applied a random forest (RF) algorithm to predict the primary sources of human clinical Salmonella Typhimurium (S. Typhimurium) and monophasic variants (monophasic S. Typhimurium) isolates. To this end, we utilised single nucleotide polymorphism diversity in the core genome MLST alleles obtained from 1,061 laboratory-confirmed human and animal S. Typhimurium and monophasic S. Typhimurium isolates as inputs into a RF model. The algorithm was used for supervised learning to classify 399 animal S. Typhimurium and monophasic S. Typhimurium isolates into one of eight distinct primary source classes comprising common livestock and pet animal species: cattle, pigs, sheep, other mammals (pets: mostly dogs and horses), broilers, layers, turkeys, and game birds (pheasants, quail, and pigeons). When applied to the training set animal isolates, model accuracy was 0.929 and kappa 0.905, whereas for the test set animal isolates, for which the primary source class information was withheld from the model, the accuracy was 0.779 and kappa 0.700. Subsequently, the model was applied to assign 662 human clinical cases to the eight primary source classes. In the dataset, 60/399 (15.0%) of the animal and 141/662 (21.3%) of the human isolates were associated with a known outbreak of S. Typhimurium definitive type (DT) 104. All but two of the 141 DT104 outbreak linked human isolates were correctly attributed by the model to the primary source classes identified as the origin of the DT104 outbreak. A model that was run without the clonal DT104 animal isolates produced largely congruent outputs (training set accuracy 0.989 and kappa 0.985; test set accuracy 0.781 and kappa 0.663). Overall, our results show that RF offers considerable promise as a suitable methodology for epidemiological tracking and source attribution for foodborne pathogens.

Minimum Inhibitory Concentrations (MICs) are the gold standard for quantitatively measuring antibiotic resistance. However, lab-based MIC determination can be time-consuming and suffers from low reproducibility, and interpretation as sensitive or resistant relies on guidelines which change over time. Genome sequencing and machine learning promise to allow in silico MIC prediction as an alternative approach which overcomes some of these difficulties, albeit the interpretation of MIC is still needed. Nevertheless, precisely how we should handle MIC data when dealing with predictive models remains unclear, since they are measured semi-quantitatively, with varying resolution, and are typically also left- and right-censored within varying ranges. We therefore investigated genome-based prediction of MICs in the pathogen Klebsiella pneumoniae using 4367 genomes with both simulated semi-quantitative traits and real MICs. As we were focused on clinical interpretation, we used interpretable rather than black-box machine learning models, namely, Elastic Net, Random Forests, and linear mixed models. Simulated traits were generated accounting for oligogenic, polygenic, and homoplastic genetic effects with different levels of heritability. Then we assessed how model prediction accuracy was affected when MICs were framed as regression and classification. Our results showed that treating the MICs differently depending on the number of concentration levels of antibiotic available was the most promising learning strategy. Specifically, to optimise both prediction accuracy and inference of the correct causal variants, we recommend considering the MICs as continuous and framing the learning problem as a regression when the number of observed antibiotic concentration levels is large, whereas with a smaller number of concentration levels they should be treated as a categorical variable and the learning problem should be framed as a classification. Our findings also underline how predictive models can be improved when prior biological knowledge is taken into account, due to the varying genetic architecture of each antibiotic resistance trait. Finally, we emphasise that incrementing the population database is pivotal for the future clinical implementation of these models to support routine machine-learning based diagnostics.

Functional genomics techniques, such as transposon insertion sequencing and RNA-sequencing, are key to studying relative differences in bacterial mutant fitness or gene expression under selective conditions. However, certain stress conditions, mutations, or antibiotics can directly interfere with DNA synthesis, resulting in systematic changes in local DNA copy numbers along the chromosome. This can lead to artifacts in sequencing-based functional genomics data when comparing antibiotic treatment to an unstressed control. Further, relative differences in gene-wise read counts may result from alterations in chromosomal replication dynamics, rather than selection or direct gene regulation. We term this artifact \"chromosomal location bias\" and implement a principled statistical approach to correct it by calculating local normalization factors along the chromosome. These normalization factors are then directly incorporated into statistical analyses using standard RNA-sequencing analysis methods without modifying the read counts themselves, preserving important information about the mean-variance relationship in the data. We illustrate the utility of this approach by generating and analyzing a ciprofloxacin-treated transposon insertion sequencing data set in Escherichia coli as a case study. We show that ciprofloxacin treatment generates chromosomal location bias in the resulting data, and we further demonstrate that failing to correct for this bias leads to false predictions of mutant drug sensitivity as measured by minimum inhibitory concentrations. We have developed an R package and user-friendly graphical Shiny application, ChromoCorrect, that detects and corrects for chromosomal bias in read count data, enabling the application of functional genomics technologies to the study of antibiotic stress.IMPORTANCEAltered gene dosage due to changes in DNA replication has been observed under a variety of stresses with a variety of experimental techniques. However, the implications of changes in gene dosage for sequencing-based functional genomics assays are rarely considered. We present a statistically principled approach to correcting for the effect of changes in gene dosage, enabling testing for differences in the fitness effects or regulation of individual genes in the presence of confounding differences in DNA copy number. We show that failing to correct for these effects can lead to incorrect predictions of resistance phenotype when applying functional genomics assays to investigate antibiotic stress, and we provide a user-friendly application to detect and correct for changes in DNA copy number.

Fundamental to effective Legionnaires\' disease outbreak control is the ability to rapidly identify the environmental source(s) of the causative agent, Legionella pneumophila. Genomics has revolutionized pathogen surveillance, but L. pneumophila has a complex ecology and population structure that can limit source inference based on standard core genome phylogenetics. Here, we present a powerful machine learning approach that assigns the geographical source of Legionnaires\' disease outbreaks more accurately than current core genome comparisons. Models were developed upon 534 L. pneumophila genome sequences, including 149 genomes linked to 20 previously reported Legionnaires\' disease outbreaks through detailed case investigations. Our classification models were developed in a cross-validation framework using only environmental L. pneumophila genomes. Assignments of clinical isolate geographic origins demonstrated high predictive sensitivity and specificity of the models, with no false positives or false negatives for 13 out of 20 outbreak groups, despite the presence of within-outbreak polyclonal population structure. Analysis of the same 534-genome panel with a conventional phylogenomic tree and a core genome multi-locus sequence type allelic distance-based classification approach revealed that our machine learning method had the highest overall classification performance-agreement with epidemiological information. Our multivariate statistical learning approach maximizes the use of genomic variation data and is thus well-suited for supporting Legionnaires\' disease outbreak investigations.IMPORTANCEIdentifying the sources of Legionnaires\' disease outbreaks is crucial for effective control. Current genomic methods, while useful, often fall short due to the complex ecology and population structure of Legionella pneumophila, the causative agent. Our study introduces a high-performing machine learning approach for more accurate geographical source attribution of Legionnaires\' disease outbreaks. Developed using cross-validation on environmental L. pneumophila genomes, our models demonstrate excellent predictive sensitivity and specificity. Importantly, this new approach outperforms traditional methods like phylogenomic trees and core genome multi-locus sequence typing, proving more efficient at leveraging genomic variation data to infer outbreak sources. Our machine learning algorithms, harnessing both core and accessory genomic variation, offer significant promise in public health settings. By enabling rapid and precise source identification in Legionnaires\' disease outbreaks, such approaches have the potential to expedite intervention efforts and curtail disease transmission.

Xanthomonas translucens, the causal agent of bacterial leaf streak disease (BLS) in cereals, is a re-emerging pathogen that is becoming increasingly destructive across the world. While BLS has caused yield losses in the past, there is anecdotal evidence that newer isolates may be more virulent. We observed that two X. translucens isolates collected from two sites in Colorado, USA, are more aggressive on current wheat and barley varieties compared to older isolates, and we hypothesize that genetic changes between recent and older isolates contribute to the differences in isolate aggressiveness. To test this, we phenotyped and genetically characterized two X. translucens isolates collected from Colorado in 2018, which we designated CO236 (from barley) and CO237 (from wheat). Using pathovar-specific phenotyping and PCR primers, we determined that CO236 belongs to pathovar translucens (Xtt) and CO237 belongs to pathovar undulosa (Xtu). We sequenced the full genomes of the isolates using Oxford Nanopore long-read sequencing, and compared their whole genomes against published X. translucens genomes. This analysis confirmed our pathovar designations for Xtt CO236 and Xtu CO237, and showed that, at the whole-genome level, there were no obvious genomic structural changes between Xtt CO236 and Xtu CO237 and other respective published pathovar genomes. Focusing on pathovar undulosa (Xtu CO237), we then compared putative type III effectors among all available Xtu isolate genomes and found that they were highly conserved. However, there were striking differences in the presence and sequence of various transcription activator-like effectors between Xtu CO237 and published undulosa genomes, which correlate with isolate virulence. Here, we explore the potential implications of the differences in these virulence factors, and provide possible explanations for the increased virulence of recently emerged isolates.

Shiga toxin-producing Escherichia coli (STEC) O157:H7 strains with the T allele in the translocated intimin receptor polymorphism (tir) 255 A > T gene associate with human disease more than strains with an A allele; however, the allele is not thought to be the direct cause of this difference. We sequenced a diverse set of STEC O157:H7 strains (26% A allele, 74% T allele) to identify linked differences that might underlie disease association. The average chromosome and pO157 plasmid size and gene content were significantly greater within the tir 255 A allele strains. Eighteen coding sequences were unique to tir 255 A allele chromosomes, and three were unique to tir 255 T allele chromosomes. There also were non-pO157 plasmids that were unique to each tir 255 allele variant. The overall average number of prophages did not differ between tir 255 allele strains; however, there were different types between the strains. Genomic and mobile element variation linked to the tir 255 polymorphism may account for the increased frequency of the T allele isolates in human disease.

Hospital bloodstream infection (BSI) caused by methicillin-resistant Staphylococcus aureus (MRSA) is a major cause of morbidity and mortality and is frequently related to invasive procedures and medically complex patients. An important feature of MRSA is the clonal structure of its population. Specific MRSA clones may differ in their pathogenic, epidemiological, and antimicrobial resistance profiles. Whole-genome sequencing is currently the most robust and discriminatory technique for tracking hypervirulent/well-adapted MRSA clones. However, it remains an expensive and time-consuming technique that requires specialized personnel. In this work, we describe a pangenome protocol, based on binary matrix (1,0) of open reading frames (ORFs), that can be used to quickly find diagnostic, apomorphic sequence mutations that can serve as biomarkers. We use this technique to create a diagnostic screen for MRSA isolates circulating in the Rio de Janeiro metropolitan area, the RdJ clone, which is prevalent in BSI. The method described here has 100% specificity and sensitivity, eliminating the need to use genomic sequencing for clonal identification. The protocol used is relatively simple and all the steps, formulas and commands used are described in this work, such that this strategy can also be used to identify other MRSA clones and even clones from other bacterial species.

Background: Advancements in DNA sequencing technology have transformed the field of bacterial genomics, allowing for faster and more cost effective chromosome level assemblies compared to a decade ago. However, transforming raw reads into a complete genome model is a significant computational challenge due to the varying quality and quantity of data obtained from different sequencing instruments, as well as intrinsic characteristics of the genome and desired analyses. To address this issue, we have developed a set of container-based pipelines using Nextflow, offering both common workflows for inexperienced users and high levels of customization for experienced ones. Their processing strategies are adaptable based on the sequencing data type, and their modularity enables the incorporation of new components to address the community\'s evolving needs. Methods: These pipelines consist of three parts: quality control, de novo genome assembly, and bacterial genome annotation. In particular, the genome annotation pipeline provides a comprehensive overview of the genome, including standard gene prediction and functional inference, as well as predictions relevant to clinical applications such as virulence and resistance gene annotation, secondary metabolite detection, prophage and plasmid prediction, and more. Results: The annotation results are presented in reports, genome browsers, and a web-based application that enables users to explore and interact with the genome annotation results. Conclusions: Overall, our user-friendly pipelines offer a seamless integration of computational tools to facilitate routine bacterial genomics research. The effectiveness of these is illustrated by examining the sequencing data of a clinical sample of Klebsiella pneumoniae.