Predicting antimicrobial resistance in Pseudomonas aeruginosa with machine learning-enabled molecular diagnostics

Unitig-centered pan-genome machine learning approach for predicting antibiotic resistance and discovering novel resistance genes in bacterial strains

In current genomic research, the widely used methods for predicting antimicrobial resistance (AMR) often rely on prior knowledge of known AMR genes or reference genomes. However, these methods have limitations, potentially resulting in imprecise predictions owing to incomplete coverage of AMR mechanisms and genetic variations. To overcome these limitations, we propose a pan-genome-based machine learning approach to advance our understanding of AMR gene repertoires and uncover possible feature sets for precise AMR classification. By building compacted de Brujin graphs (cDBGs) from thousands of genomes and collecting the presence/absence patterns of unique sequences (unitigs) for Pseudomonas aeruginosa, we determined that using machine learning models on unitig-centered pan-genomes showed significant promise for accurately predicting the antibiotic resistance or susceptibility of microbial strains. Applying a feature-selection-based machine learning algorithm led to satisfactory predictive performance for the training dataset (with an area under the receiver operating characteristic curve (AUC) of > 0.929) and an independent validation dataset (AUC, approximately 0.77). Furthermore, the selected unitigs revealed previously unidentified resistance genes, allowing for the expansion of the resistance gene repertoire to those that have not previously been described in the literature on antibiotic resistance. These results demonstrate that our proposed unitig-based pan-genome feature set was effective in constructing machine learning predictors that could accurately identify AMR pathogens. Gene sets extracted using this approach may offer valuable insights into expanding known AMR genes and forming new hypotheses to uncover the underlying mechanisms of bacterial AMR.

Variation in the response to antibiotics and life-history across the major Pseudomonas aeruginosa clone type (mPact) panel

Pseudomonas aeruginosa is a ubiquitous, opportunistic human pathogen. Since it often expresses multidrug resistance, new treatment options are urgently required. Such new treatments are usually assessed with one of the canonical laboratory strains, PAO1 or PA14. However, these two strains are unlikely representative of the strains infecting patients, because they have adapted to laboratory conditions and do not capture the enormous genomic diversity of the species. Here, we characterized the major P. aeruginosa clone type (mPact) panel. This panel consists of 20 strains, which reflect the species' genomic diversity, cover all major clone types, and have both patient and environmental origins. We found significant strain variation in distinct responses toward antibiotics and general growth characteristics. Only few of the measured traits are related, suggesting independent trait optimization across strains. High resistance levels were only identified for clinical mPact isolates and could be linked to known antimicrobial resistance (AMR) genes. One strain, H01, produced highly unstable AMR combined with reduced growth under drug-free conditions, indicating an evolutionary cost to resistance. The expression of microcolonies was common among strains, especially for strain H15, which also showed reduced growth, possibly indicating another type of evolutionary trade-off. By linking isolation source, growth, and virulence to life history traits, we further identified specific adaptive strategies for individual mPact strains toward either host processes or degradation pathways. Overall, the mPact panel provides a reasonably sized set of distinct strains, enabling in-depth analysis of new treatment designs or evolutionary dynamics in consideration of the species' genomic diversity.

IMPORTANCE

New treatment strategies are urgently needed for high-risk pathogens such as the opportunistic and often multidrug-resistant pathogen Pseudomonas aeruginosa. Here, we characterize the major P. aeruginosa clone type (mPact) panel. It consists of 20 strains with different origins that cover the major clone types of the species as well as its genomic diversity. This mPact panel shows significant variation in (i) resistance against distinct antibiotics, including several last resort antibiotics; (ii) related traits associated with the response to antibiotics; and (iii) general growth characteristics. We further developed a novel approach that integrates information on resistance, growth, virulence, and life-history characteristics, allowing us to demonstrate the presence of distinct adaptive strategies of the strains that focus either on host interaction or resource processing. In conclusion, the mPact panel provides a manageable number of representative strains for this important pathogen for further in-depth analyses of treatment options and evolutionary dynamics.

Keeping up with the pathogens: improved antimicrobial resistance detection and prediction from Pseudomonas aeruginosa genomes

BACKGROUND

Antimicrobial resistance (AMR) is an intensifying threat that requires urgent mitigation to avoid a post-antibiotic era. Pseudomonas aeruginosa represents one of the greatest AMR concerns due to increasing multi- and pan-drug resistance rates. Shotgun sequencing is gaining traction for in silico AMR profiling due to its unambiguity and transferability; however, accurate and comprehensive AMR prediction from P. aeruginosa genomes remains an unsolved problem.

METHODS

We first curated the most comprehensive database yet of known P. aeruginosa AMR variants. Next, we performed comparative genomics and microbial genome-wide association study analysis across a Global isolate Dataset (n = 1877) with paired antimicrobial phenotype and genomic data to identify novel AMR variants. Finally, the performance of our P. aeruginosa AMR database, implemented in our AMR detection and prediction tool, ARDaP, was compared with three previously published in silico AMR gene detection or phenotype prediction tools-abritAMR, AMRFinderPlus, ResFinder-across both the Global Dataset and an analysis-naïve Validation Dataset (n = 102).

RESULTS

Our AMR database comprises 3639 mobile AMR genes and 728 chromosomal variants, including 75 previously unreported chromosomal AMR variants, 10 variants associated with unusual antimicrobial susceptibility, and 281 chromosomal variants that we show are unlikely to confer AMR. Our pipeline achieved a genotype-phenotype balanced accuracy (bACC) of 85% and 81% across 10 clinically relevant antibiotics when tested against the Global and Validation Datasets, respectively, vs. just 56% and 54% with abritAMR, 58% and 54% with AMRFinderPlus, and 60% and 53% with ResFinder. ARDaP's superior performance was predominantly due to the inclusion of chromosomal AMR variants, which are generally not identified with most AMR identification tools.

CONCLUSIONS

Our ARDaP software and associated AMR variant database provides an accurate tool for predicting AMR phenotypes in P. aeruginosa, far surpassing the performance of current tools. Implementation of ARDaP for routine AMR prediction from P. aeruginosa genomes and metagenomes will improve AMR identification, addressing a critical facet in combatting this treatment-refractory pathogen. However, knowledge gaps remain in our understanding of the P. aeruginosa resistome, particularly the basis of colistin AMR.

Direct prediction of antimicrobial resistance in Pseudomonas aeruginosa by metagenomic next-generation sequencing

Objective

Pseudomonas aeruginosa has strong drug resistance and can tolerate a variety of antibiotics, which is a major problem in the management of antibiotic-resistant infections. Direct prediction of multi-drug resistance (MDR) resistance phenotypes of P. aeruginosa isolates and clinical samples by genotype is helpful for timely antibiotic treatment.

Methods

In the study, whole genome sequencing (WGS) data of 494 P. aeruginosa isolates were used to screen key anti-microbial resistance (AMR)-associated genes related to imipenem (IPM), meropenem (MEM), piperacillin/tazobactam (TZP), and levofloxacin (LVFX) resistance in P. aeruginosa by comparing genes with copy number differences between resistance and sensitive strains. Subsequently, for the direct prediction of the resistance of P. aeruginosa to four antibiotics by the AMR-associated features screened, we collected 74 P. aeruginosa positive sputum samples to sequence by metagenomics next-generation sequencing (mNGS), of which 1 sample with low quality was eliminated. Then, we constructed the resistance prediction model.

Results

We identified 93, 88, 80, 140 AMR-associated features for IPM, MEM, TZP, and LVFX resistance in P. aeruginosa. The relative abundance of AMR-associated genes was obtained by matching mNGS and WGS data. The top 20 features with importance degree for IPM, MEM, TZP, and LVFX resistance were used to model, respectively. Then, we used the random forest algorithm to construct resistance prediction models of P. aeruginosa, in which the areas under the curves of the IPM, MEM, TZP, and LVFX resistance prediction models were all greater than 0.8, suggesting these resistance prediction models had good performance.

Conclusion

In summary, mNGS can predict the resistance of P. aeruginosa by directly detecting AMR-associated genes, which provides a reference for rapid clinical detection of drug resistance of pathogenic bacteria.

Machine learning-based antibiotic resistance prediction models: An updated systematic review and meta-analysis

BACKGROUND

The widespread use of antibiotics has led to a gradual adaptation of bacteria to these drugs, diminishing the effectiveness of treatments.

OBJECTIVE

To comprehensively assess the research progress of antibiotic resistance prediction models based on machine learning (ML) algorithms, providing the latest quantitative analysis and methodological evaluation.

METHODS

Relevant literature was systematically retrieved from databases, including PubMed, Embase and the Cochrane Library, from inception up to December 2023. Studies meeting predefined criteria were selected for inclusion. The prediction model risk of bias assessment tool was employed for methodological quality assessment, and a random-effects model was utilised for meta-analysis.

RESULTS

The systematic review included a total of 22 studies with a combined sample size of 43,628; 10 studies were ultimately included in the meta-analysis. Commonly used ML algorithms included random forest, decision trees and neural networks. Frequently utilised predictive variables encompassed demographics, drug use history and underlying diseases. The overall sensitivity was 0.57 (95% CI: 0.42-0.70; p< 0.001; I2= 99.7%), the specificity was 0.95 (95% CI: 0.79-0.99; p< 0.001; I2 = 99.9%), the positive likelihood ratio was 10.7 (95% CI: 2.9-39.5), the negative likelihood ratio was 0.46 (95% CI: 0.34-0.61), the diagnostic odds ratio was 23 (95% CI: 7-81) and the area under the receiver operating characteristic curve was 0.78 (95% CI: 0.74-0.81; p< 0.001), indicating a good discriminative ability of ML models for antibiotic resistance. However, methodological assessment and funnel plots suggested a high risk of bias and publication bias in the included studies.

CONCLUSION

This meta-analysis provides a current and comprehensive evaluation of ML models for predicting antibiotic resistance, emphasising their potential application in clinical practice. Nevertheless, stringent research design and reporting are warranted to enhance the quality and credibility of future studies. Future research should focus on methodological innovation and incorporate more high-quality studies to further advance this field.

Tackling the Antimicrobial Resistance "Pandemic" with Machine Learning Tools: A Summary of Available Evidence

Antimicrobial resistance is recognised as one of the top threats healthcare is bound to face in the future. There have been various attempts to preserve the efficacy of existing antimicrobials, develop new and efficient antimicrobials, manage infections with multi-drug resistant strains, and improve patient outcomes, resulting in a growing mass of routinely available data, including electronic health records and microbiological information that can be employed to develop individualised antimicrobial stewardship. Machine learning methods have been developed to predict antimicrobial resistance from whole-genome sequencing data, forecast medication susceptibility, recognise epidemic patterns for surveillance purposes, or propose new antibacterial treatments and accelerate scientific discovery. Unfortunately, there is an evident gap between the number of machine learning applications in science and the effective implementation of these systems. This narrative review highlights some of the outstanding opportunities that machine learning offers when applied in research related to antimicrobial resistance. In the future, machine learning tools may prove to be superbugs' kryptonite. This review aims to provide an overview of available publications to aid researchers that are looking to expand their work with new approaches and to acquaint them with the current application of machine learning techniques in this field.

Antimicrobial resistance prediction by clinical metagenomics in pediatric severe pneumonia patients

BACKGROUND

Antimicrobial resistance (AMR) is a major threat to children's health, particularly in respiratory infections. Accurate identification of pathogens and AMR is crucial for targeted antibiotic treatment. Metagenomic next-generation sequencing (mNGS) shows promise in directly detecting microorganisms and resistance genes in clinical samples. However, the accuracy of AMR prediction through mNGS testing needs further investigation for practical clinical decision-making.

METHODS

We aimed to evaluate the performance of mNGS in predicting AMR for severe pneumonia in pediatric patients. We conducted a retrospective analysis at a tertiary hospital from May 2022 to May 2023. Simultaneous mNGS and culture were performed on bronchoalveolar lavage fluid samples obtained from pediatric patients with severe pneumonia. By comparing the results of mNGS detection of microorganisms and antibiotic resistance genes with those of culture, sensitivity, specificity, positive predictive value, and negative predictive value were calculated.

RESULTS

mNGS detected bacterial in 71.7% cases (86/120), significantly higher than culture (58/120, 48.3%). Compared to culture, mNGS demonstrated a sensitivity of 96.6% and a specificity of 51.6% in detecting pathogenic microorganisms. Phenotypic susceptibility testing (PST) of 19 antibiotics revealed significant variations in antibiotics resistance rates among different bacteria. Sensitivity prediction of mNGS for carbapenem resistance was higher than penicillins and cephalosporin (67.74% vs. 28.57%, 46.15%), while specificity showed no significant difference (85.71%, 75.00%, 75.00%). mNGS also showed a high sensitivity of 94.74% in predicting carbapenem resistance in Acinetobacter baumannii.

CONCLUSIONS

mNGS exhibits variable predictive performance among different pathogens and antibiotics, indicating its potential as a supplementary tool to conventional PST. However, mNGS currently cannot replace conventional PST.

Assessing computational predictions of antimicrobial resistance phenotypes from microbial genomes

The advent of rapid whole-genome sequencing has created new opportunities for computational prediction of antimicrobial resistance (AMR) phenotypes from genomic data. Both rule-based and machine learning (ML) approaches have been explored for this task, but systematic benchmarking is still needed. Here, we evaluated four state-of-the-art ML methods (Kover, PhenotypeSeeker, Seq2Geno2Pheno and Aytan-Aktug), an ML baseline and the rule-based ResFinder by training and testing each of them across 78 species-antibiotic datasets, using a rigorous benchmarking workflow that integrates three evaluation approaches, each paired with three distinct sample splitting methods. Our analysis revealed considerable variation in the performance across techniques and datasets. Whereas ML methods generally excelled for closely related strains, ResFinder excelled for handling divergent genomes. Overall, Kover most frequently ranked top among the ML approaches, followed by PhenotypeSeeker and Seq2Geno2Pheno. AMR phenotypes for antibiotic classes such as macrolides and sulfonamides were predicted with the highest accuracies. The quality of predictions varied substantially across species-antibiotic combinations, particularly for beta-lactams; across species, resistance phenotyping of the beta-lactams compound, aztreonam, amoxicillin/clavulanic acid, cefoxitin, ceftazidime and piperacillin/tazobactam, alongside tetracyclines demonstrated more variable performance than the other benchmarked antibiotics. By organism, Campylobacter jejuni and Enterococcus faecium phenotypes were more robustly predicted than those of Escherichia coli, Staphylococcus aureus, Salmonella enterica, Neisseria gonorrhoeae, Klebsiella pneumoniae, Pseudomonas aeruginosa, Acinetobacter baumannii, Streptococcus pneumoniae and Mycobacterium tuberculosis. In addition, our study provides software recommendations for each species-antibiotic combination. It furthermore highlights the need for optimization for robust clinical applications, particularly for strains that diverge substantially from those used for training.

tRNA epitranscriptome determines pathogenicity of the opportunistic pathogen Pseudomonas aeruginosa

The success of bacterial pathogens depends on the coordinated expression of virulence determinants. Regulatory circuits that drive pathogenesis are complex, multilayered, and incompletely understood. Here, we reveal that alterations in tRNA modifications define pathogenic phenotypes in the opportunistic pathogen Pseudomonas aeruginosa. We demonstrate that the enzymatic activity of GidA leads to the introduction of a carboxymethylaminomethyl modification in selected tRNAs. Modifications at the wobble uridine base (cmnm5U34) of the anticodon drives translation of transcripts containing rare codons. Specifically, in P. aeruginosa the presence of GidA-dependent tRNA modifications modulates expression of genes encoding virulence regulators, leading to a cellular proteomic shift toward pathogenic and well-adapted physiological states. Our approach of profiling the consequences of chemical tRNA modifications is general in concept. It provides a paradigm of how environmentally driven tRNA modifications govern gene expression programs and regulate phenotypic outcomes responsible for bacterial adaption to challenging habitats prevailing in the host niche.

Working with the Electronic Health Record and Laboratory Information System to Maximize Ordering and Reporting of Molecular Microbiology Results

Molecular microbiology assays have a higher cost of testing compared to traditional methods and need to be utilized appropriately. Results from these assays may also require interpretation and appropriate follow-up. Electronic tools available in the electronic health record and laboratory information system can be deployed both preanalytically and postanalytically to influence ordering behaviors and positively impact diagnostic stewardship. Next generation technologies, such as machine learning and artificial intelligence, have the potential to expand upon the capabilities currently available and warrant additional study and development but also require regulation around their use in health care.

Predictive Modeling of Phenotypic Antimicrobial Susceptibility of Selected Beta-Lactam Antimicrobials from Beta-Lactamase Resistance Genes

The outcome of bacterial infection management relies on prompt diagnosis and effective treatment, but conventional antimicrobial susceptibility testing can be slow and labor-intensive. Therefore, this study aims to predict phenotypic antimicrobial susceptibility of selected beta-lactam antimicrobials in the bacteria of the family Enterobacteriaceae from different beta-lactamase resistance genotypes. Using human datasets extracted from the Antimicrobial Testing Leadership and Surveillance (ATLAS) program conducted by Pfizer and retail meat datasets from the National Antimicrobial Resistance Monitoring System for Enteric Bacteria (NARMS), we used a robust or weighted least square multivariable linear regression modeling framework to explore the relationship between antimicrobial susceptibility data of beta-lactam antimicrobials and different types of beta-lactamase resistance genes. In humans, in the presence of the blaCTX-M-1, blaCTX-M-2, blaCTX-M-8/25, and blaCTX-M-9 groups, MICs of cephalosporins significantly increased by values between 0.34-3.07 μg/mL, however, the MICs of carbapenem significantly decreased by values between 0.81-0.87 μg/mL. In the presence of carbapenemase genes (blaKPC, blaNDM, blaIMP, and blaVIM), the MICs of cephalosporin antimicrobials significantly increased by values between 1.06-5.77 μg/mL, while the MICs of carbapenem antimicrobials significantly increased by values between 5.39-67.38 μg/mL. In retail meat, MIC of ceftriaxone increased significantly in the presence of blaCMY-2, blaCTX-M-1, blaCTX-M-55, blaCTX-M-65, and blaSHV-2 by 55.16 μg/mL, 222.70 μg/mL, 250.81 μg/mL, 204.89 μg/mL, and 31.51 μg/mL respectively. MIC of cefoxitin increased significantly in the presence of blaCTX-M-65 and blaTEM-1 by 1.57 μg/mL and 1.04 μg/mL respectively. In the presence of blaCMY-2, MIC of cefoxitin increased by an average of 8.66 μg/mL over 17 years. Compared to E. coli isolates, MIC of cefoxitin in Salmonella enterica isolates decreased significantly by 0.67 μg/mL. On the other hand, MIC of ceftiofur increased in the presence of blaCTX-M-1, blaCTX-M-65, blaSHV-2, and blaTEM-1 by 8.82 μg/mL, 9.11 μg/mL, 8.18 μg/mL, and 1.04 μg/mL respectively. In the presence of blaCMY-2, MIC of ceftiofur increased by an average of 10.20 μg/mL over 14 years. The ability to predict antimicrobial susceptibility of beta-lactam antimicrobials directly from beta-lactamase resistance genes may help reduce the reliance on routine phenotypic testing with higher turnaround times in diagnostic, therapeutic, and surveillance of antimicrobial-resistant bacteria of the family Enterobacteriaceae.

Antimicrobial resistance crisis: could artificial intelligence be the solution?

Antimicrobial resistance is a global public health threat, and the World Health Organization (WHO) has announced a priority list of the most threatening pathogens against which novel antibiotics need to be developed. The discovery and introduction of novel antibiotics are time-consuming and expensive. According to WHO's report of antibacterial agents in clinical development, only 18 novel antibiotics have been approved since 2014. Therefore, novel antibiotics are critically needed. Artificial intelligence (AI) has been rapidly applied to drug development since its recent technical breakthrough and has dramatically improved the efficiency of the discovery of novel antibiotics. Here, we first summarized recently marketed novel antibiotics, and antibiotic candidates in clinical development. In addition, we systematically reviewed the involvement of AI in antibacterial drug development and utilization, including small molecules, antimicrobial peptides, phage therapy, essential oils, as well as resistance mechanism prediction, and antibiotic stewardship.

Artificial intelligence and machine learning in clinical pharmacological research

BACKGROUND

Clinical pharmacology research has always involved computational analysis. With the abundance of drug-related data available, the integration of artificial intelligence (AI) and machine learning (ML) methods has emerged as a promising way to enhance clinical pharmacology research.

METHODS

Based on an accepted definition of clinical pharmacology as a field of research dealing with all aspects of drug-human interactions, the analysis included publications from institutes specializing in clinical pharmacology. Research topics and the most used machine learning methods in clinical pharmacology were retrieved from the PubMed database and summarized.

RESULTS

ML was identified in 674 publications attributed to clinical pharmacology research, with a significant increase in publication activity over the last decade. Notable research topics addressed by ML/AI included Covid-19-related clinical pharmacology research, clinical neuropharmacology, drug safety and risk assessment, clinical pharmacology related to cancer research, and antimicrobial and antiviral research unrelated to Covid-19. In terms of ML methods, neural networks, random forests, and support vector machines were frequently mentioned in the abstracts of the retrieved papers.

CONCLUSIONS

ML, and AI in general, is increasingly being used in various research areas within clinical pharmacology. This report presents specific examples of applications and highlights the most used ML methods.

Comparative genomics reveals the correlations of stress response genes and bacteriophages in developing antibiotic resistance of Staphylococcus saprophyticus

IMPORTANCE

Staphylococcus saprophyticus is the second most common bacteria associated with urinary tract infections (UTIs) in women. The antimicrobial treatment regimen for uncomplicated UTI is normally nitrofurantoin, trimethoprim-sulfamethoxazole (TMP-SMX), or a fluoroquinolone without routine susceptibility testing of S. saprophyticus recovered from urine specimens. However, TMP-SMX-resistant S. saprophyticus has been detected recently in UTI patients, as well as in our cohort. Herein, we investigated the understudied resistance patterns of this pathogenic species by linking genomic antibiotic resistance gene (ARG) content to susceptibility phenotypes. We describe ARG associations with known and novel SCCmec configurations as well as phage elements in S. saprophyticus, which may serve as intervention or diagnostic targets to limit resistance transmission. Our analyses yielded a comprehensive database of phenotypic data associated with the ARG sequence in clinical S. saprophyticus isolates, which will be crucial for resistance surveillance and prediction to enable precise diagnosis and effective treatment of S. saprophyticus UTIs.

Machine learning and phylogenetic analysis allow for predicting antibiotic resistance in M. tuberculosis

BACKGROUND

Antimicrobial resistance (AMR) poses a significant global health threat, and an accurate prediction of bacterial resistance patterns is critical for effective treatment and control strategies. In recent years, machine learning (ML) approaches have emerged as powerful tools for analyzing large-scale bacterial AMR data. However, ML methods often ignore evolutionary relationships among bacterial strains, which can greatly impact performance of the ML methods, especially if resistance-associated features are attempted to be detected. Genome-wide association studies (GWAS) methods like linear mixed models accounts for the evolutionary relationships in bacteria, but they uncover only highly significant variants which have already been reported in literature.

RESULTS

In this work, we introduce a novel phylogeny-related parallelism score (PRPS), which measures whether a certain feature is correlated with the population structure of a set of samples. We demonstrate that PRPS can be used, in combination with SVM- and random forest-based models, to reduce the number of features in the analysis, while simultaneously increasing models' performance. We applied our pipeline to publicly available AMR data from PATRIC database for Mycobacterium tuberculosis against six common antibiotics.

CONCLUSIONS

Using our pipeline, we re-discovered known resistance-associated mutations as well as new candidate mutations which can be related to resistance and not previously reported in the literature. We demonstrated that taking into account phylogenetic relationships not only improves the model performance, but also yields more biologically relevant predicted most contributing resistance markers.

Direct prediction of carbapenem resistance in Pseudomonas aeruginosa by whole genome sequencing and metagenomic sequencing

Carbapenem resistance is a major concern in the management of antibiotic-resistant Pseudomonas aeruginosa infections. The direct prediction of carbapenem-resistant phenotype from genotype in P. aeruginosa isolates and clinical samples would promote timely antibiotic therapy. The complex carbapenem resistance mechanism and the high prevalence of variant-driven carbapenem resistance in P. aeruginosa make it challenging to predict the carbapenem-resistant phenotype through the genotype. In this study, using whole genome sequencing (WGS) data of 1,622 P. aeruginosa isolates followed by machine learning, we screened 16 and 31 key gene features associated with imipenem (IPM) and meropenem (MEM) resistance in P. aeruginosa, including oprD(HIGH), and constructed the resistance prediction models. The areas under the curves of the IPM and MEM resistance prediction models were 0.906 and 0.925, respectively. For the direct prediction of carbapenem resistance in P. aeruginosa from clinical samples by the key gene features selected and prediction models constructed, 72 P. aeruginosa-positive sputum samples were collected and sequenced by metagenomic sequencing (MGS) based on next-generation sequencing (NGS) or Oxford Nanopore Technology (ONT). The prediction applicability of MGS based on NGS outperformed that of MGS based on ONT. In 72 P. aeruginosa-positive sputum samples, 65.0% (26/40) of IPM-insensitive and 63.2% (24/38) of MEM-insensitive P. aeruginosa were directly predicted by NGS-based MGS with positive predictive values of 0.897 and 0.889, respectively. By the direct detection of the key gene features associated with carbapenem resistance of P. aeruginosa, the carbapenem resistance of P. aeruginosa could be directly predicted from cultured isolates by WGS or from clinical samples by NGS-based MGS, which could assist the timely treatment and surveillance of carbapenem-resistant P. aeruginosa.

Rapid AMR prediction in Pseudomonas aeruginosa combining MALDI-TOF MS with DNN model

BACKGROUND

Pseudomonas aeruginosa is a significant clinical pathogen that poses a substantial threat due to its extensive drug resistance. The rapid and precise identification of this resistance is crucial for effective clinical treatment. Although matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) has been used for antibiotic susceptibility differentiation of some bacteria in recent years, the genetic diversity of P. aeruginosa complicates population analysis. Rapid identification of antimicrobial resistance (AMR) in P. aeruginosa based on a large amount of MALDI-TOF-MS data has not yet been reported. In this study, we employed publicly available datasets for P. aeruginosa, which contain data on bacterial resistance and MALDI-TOF-MS spectra. We introduced a deep neural network model, synergized with a strategic sampling approach (SMOTEENN) to construct a predictive framework for AMR of three widely used antibiotics.

RESULTS

The framework achieved area under the curve values of 90%, 85%, and 77% for Tobramycin, Cefepime, and Meropenem, respectively, surpassing conventional classifiers. Notably, random forest algorithm was used to assess the significance of features and post-hoc analysis was conducted on the top 10 features using Cohen's d. This analysis revealed moderate effect sizes (d = 0.5-0.8) in Tobramycin and Cefepime models. Finally, putative AMR biomarkers were identified in this study.

CONCLUSIONS

This work presented an AMR prediction tool specifically designed for P. aeruginosa, which offers a hopeful pathway for clinical decision-making.

Challenges and opportunities in sharing microbiome data and analyses

Microbiome data, metadata and analytical workflows have become 'big' in terms of volume and complexity. Although the infrastructure and technologies to share data have been established, the interdisciplinary and multi-omic nature of the field can make resources difficult to identify and use. Following best practices for data deposition requires substantial effort, with sometimes little obvious reward. Gaps remain where microbiome-specific resources for data sharing or reproducibility do not yet exist. We outline available best practices, challenges to their adoption and opportunities in data sharing in microbiome research. We showcase examples of best practices and advocate for their enforcement and incentivization for data sharing. This includes recognition of data curation and sharing endeavours by individuals, institutions, journals and funders. Opportunities for progress include enabling microbiome-specific databases to incorporate future methods for data analysis, integration and reuse.

Predictive modeling of antibiotic eradication therapy success for new-onset Pseudomonas aeruginosa pulmonary infections in children with cystic fibrosis

Chronic Pseudomonas aeruginosa (Pa) lung infections are the leading cause of mortality among cystic fibrosis (CF) patients; therefore, the eradication of new-onset Pa lung infections is an important therapeutic goal that can have long-term health benefits. The use of early antibiotic eradication therapy (AET) has been shown to clear the majority of new-onset Pa infections, and it is hoped that identifying the underlying basis for AET failure will further improve treatment outcomes. Here we generated machine learning models to predict AET outcomes based on pathogen genomic data. We used a nested cross validation design, population structure control, and recursive feature selection to improve model performance and showed that incorporating population structure control was crucial for improving model interpretation and generalizability. Our best model, controlling for population structure and using only 30 recursively selected features, had an area under the curve of 0.87 for a holdout test dataset. The top-ranked features were generally associated with motility, adhesion, and biofilm formation.

Fighting Pseudomonas aeruginosa Infections: Antibacterial and Antibiofilm Activity of D-Q53 CecB, a Synthetic Analog of a Silkworm Natural Cecropin B Variant

Pseudomonas aeruginosa is an opportunistic Gram-negative bacterium responsible for severe nosocomial infections and is considered a critical pulmonary pathogen for both immunocompromised and cystic fibrosis patients. Planktonic cells of P. aeruginosa possess intrinsic and acquired resistances, inactivating several classes of conventional antibiotics. Additionally, this bacterium can grow, forming biofilms, and complex structures, further hampering the action of multiple antibiotics. Here, we report the biological properties of D-Q53 CecB, an all-D enantiomer of the silkworm natural peptide Q53 CecB. Compared to the L-variant, D-Q53 CecB was resistant to in vitro degradation by humans and P. aeruginosa elastases and showed an enhanced bactericidal activity against P. aeruginosa planktonic bacteria. D-Q53 CecB was thermostable and maintained its antimicrobial activity at high salt concentrations and in the presence of divalent cations or fetal-bovine serum, although at reduced levels. Against different types of human cells, D-Q53 CecB showed cytotoxic phenomena at concentrations several folds higher compared to those active against P. aeruginosa. When L- and D-Q53 CecB were compared for their antibiofilm properties, both peptides were active in inhibiting biofilm formation. However, the D-enantiomer was extremely effective in inducing biofilm degradation, suggesting this peptide as a favorable candidate in an anti-Pseudomonas therapy.

Leveraging artificial intelligence in the fight against infectious diseases

Despite advances in molecular biology, genetics, computation, and medicinal chemistry, infectious disease remains an ominous threat to public health. Addressing the challenges posed by pathogen outbreaks, pandemics, and antimicrobial resistance will require concerted interdisciplinary efforts. In conjunction with systems and synthetic biology, artificial intelligence (AI) is now leading to rapid progress, expanding anti-infective drug discovery, enhancing our understanding of infection biology, and accelerating the development of diagnostics. In this Review, we discuss approaches for detecting, treating, and understanding infectious diseases, underscoring the progress supported by AI in each case. We suggest future applications of AI and how it might be harnessed to help control infectious disease outbreaks and pandemics.

Application of machine learning in the prevention of periprosthetic joint infection following total knee arthroplasty: a systematic review

BACKGROUND

Machine learning is a promising and powerful technology with increasing use in orthopedics. Periprosthetic joint infection following total knee arthroplasty results in increased morbidity and mortality. This systematic review investigated the use of machine learning in preventing periprosthetic joint infection.

METHODS

A systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. PubMed was searched in November 2022. All studies that investigated the clinical applications of machine learning in the prevention of periprosthetic joint infection following total knee arthroplasty were included. Non-English studies, studies with no full text available, studies focusing on non-clinical applications of machine learning, reviews and meta-analyses were excluded. For each included study, its characteristics, machine learning applications, algorithms, statistical performances, strengths and limitations were summarized. Limitations of the current machine learning applications and the studies, including their 'black box' nature, overfitting, the requirement of a large dataset, the lack of external validation, and their retrospective nature were identified.

RESULTS

Eleven studies were included in the final analysis. Machine learning applications in the prevention of periprosthetic joint infection were divided into four categories: prediction, diagnosis, antibiotic application and prognosis.

CONCLUSION

Machine learning may be a favorable alternative to manual methods in the prevention of periprosthetic joint infection following total knee arthroplasty. It aids in preoperative health optimization, preoperative surgical planning, the early diagnosis of infection, the early application of suitable antibiotics, and the prediction of clinical outcomes. Future research is warranted to resolve the current limitations and bring machine learning into clinical settings.

Using bacterial pan-genome-based feature selection approach to improve the prediction of minimum inhibitory concentration (MIC)

Background: Predicting the resistance profiles of antimicrobial resistance (AMR) pathogens is becoming more and more important in treating infectious diseases. Various attempts have been made to build machine learning models to classify resistant or susceptible pathogens based on either known antimicrobial resistance genes or the entire gene set. However, the phenotypic annotations are translated from minimum inhibitory concentration (MIC), which is the lowest concentration of antibiotic drugs in inhibiting certain pathogenic strains. Since the MIC breakpoints that classify a strain to be resistant or susceptible to specific antibiotic drug may be revised by governing institutes, we refrained from translating these MIC values into the categories "susceptible" or "resistant" but instead attempted to predict the MIC values using machine learning approaches. Results: By applying a machine learning feature selection approach on a Salmonella enterica pan-genome, in which the protein sequences were clustered to identify highly similar gene families, we showed that the selected features (genes) performed better than known AMR genes, and that models built on the selected genes achieved very accurate MIC prediction. Functional analysis revealed that about half of the selected genes were annotated as hypothetical proteins (i.e., with unknown functional roles), and that only a small portion of known AMR genes were among the selected genes, indicating that applying feature selection on the entire gene set has the potential of uncovering novel genes that may be associated with and may contribute to pathogenic antimicrobial resistances. Conclusion: The application of the pan-genome-based machine learning approach was indeed capable of predicting MIC values with very high accuracy. The feature selection process may also identify novel AMR genes for inferring bacterial antimicrobial resistance phenotypes.

Identification of a Specific Biomarker of Acinetobacter baumannii Global Clone 1 by Machine Learning and PCR Related to Metabolic Fitness of ESKAPE Pathogens

Since the emergence of high-risk clones worldwide, constant investigations have been undertaken to comprehend the molecular basis that led to their prevalent dissemination in nosocomial settings over time. So far, the complex and multifactorial genetic traits of this type of epidemic clones have allowed only the identification of biomarkers with low specificity. A machine learning algorithm was able to recognize unequivocally a biomarker for early and accurate detection of Acinetobacter baumannii global clone 1 (GC1), one of the most disseminated high-risk clones. A support vector machine model identified the U1 sequence with a length of 367 nucleotides that matched a fragment of the moaCB gene, which encodes the molybdenum cofactor biosynthesis C and B proteins. U1 differentiates specifically between A. baumannii GC1 and non-GC1 strains, becoming a suitable biomarker capable of being translated into clinical settings as a molecular typing method for early diagnosis based on PCR as shown here. Since the metabolic pathways of Mo enzymes have been recognized as putative therapeutic targets for ESKAPE (Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacter species) pathogens, our findings highlight that machine learning can also be useful in knowledge gaps of high-risk clones and provides noteworthy support to the literature to identify relevant nosocomial biomarkers for other multidrug-resistant high-risk clones. IMPORTANCE A. baumannii GC1 is an important high-risk clone that rapidly develops extreme drug resistance in the nosocomial niche. Furthermore, several strains have been identified worldwide in environmental samples, exacerbating the risk of human interactions. Early diagnosis is mandatory to limit its dissemination and to outline appropriate antibiotic stewardship schedules. A region with a length of 367 bp (U1) within the moaCB gene that is not subjected to lateral genetic transfer or to antibiotic pressures was successfully found by a support vector machine model that predicts A. baumannii GC1 strains. At the same time, research on the group of Mo enzymes proposed this metabolic pathway related to the superbug's metabolism as a potential future drug target site for ESKAPE pathogens due to its central role in bacterial fitness during infection. These findings confirm that machine learning used for the identification of biomarkers of high-risk lineages can also serve to identify putative novel therapeutic target sites.

Genomic Surveillance of Salmonella from the Comunitat Valenciana (Spain)

Salmonella enterica subspecies enterica is one of the most important foodborne pathogens and the causative agent of salmonellosis, which affects both humans and animals producing numerous infections every year. The study and understanding of its epidemiology are key to monitoring and controlling these bacteria. With the development of whole-genome sequencing (WGS) technologies, surveillance based on traditional serotyping and phenotypic tests of resistance is being replaced by genomic surveillance. To introduce WGS as a routine methodology for the surveillance of food-borne Salmonella in the region, we applied this technology to analyze a set of 141 S. enterica isolates obtained from various food sources between 2010 and 2017 in the Comunitat Valenciana (Spain). For this, we performed an evaluation of the most relevant Salmonella typing methods, serotyping and sequence typing, using both traditional and in silico approaches. We extended the use of WGS to detect antimicrobial resistance determinants and predicted minimum inhibitory concentrations (MICs). Finally, to understand possible contaminant sources in this region and their relationship to antimicrobial resistance (AMR), we performed cluster detection combining single-nucleotide polymorphism (SNP) pairwise distances and phylogenetic and epidemiological data. The results of in silico serotyping with WGS data were highly congruent with those of serological analyses (98.5% concordance). Multi-locus sequence typing (MLST) profiles obtained with WGS information were also highly congruent with the sequence type (ST) assignment based on Sanger sequencing (91.9% coincidence). In silico identification of antimicrobial resistance determinants and minimum inhibitory concentrations revealed a high number of resistance genes and possible resistant isolates. A combined phylogenetic and epidemiological analysis with complete genome sequences revealed relationships among isolates indicative of possible common sources for isolates with separate sampling in time and space that had not been detected from epidemiological information. As a result, we demonstrate the usefulness of WGS and in silico methods to obtain an improved characterization of S. enterica enterica isolates, allowing better surveillance of the pathogen in food products and in potential environmental and clinical samples of related interest.

Increased growth temperature and vitamin B12 supplementation reduces the lag time for rapid pathogen identification in BHI agar and blood cultures

Background: Rapid diagnostics of pathogens is essential to prescribe appropriate antibiotic therapy. The current methods for pathogen detection require the bacteria to grow in a culture medium, which is time-consuming. This increases the mortality rate and global burden of antimicrobial resistance. Culture-free detection methods are still under development and are not common in the clinical routine. Therefore, decreasing the culture time for accurately detecting infection and resistance is vital for diagnosis. Methods: This study investigated easy-to-implement factors (in a minimal laboratory set-up), including inoculum size, incubation temperature, and additional supplementation (e.g., vitamin B12 and trace metals), that can significantly reduce the bacterial lag time (tlag). These factors were arranged in simple two-level factorial designs using Gram-positive cocci (Staphylococcus aureus), Gram-positive bacilli (Bacillus subtilis), and Gram-negative bacilli (Escherichia coli and Pseudomonas aeruginosa) bacteria, including clinical isolates with known antimicrobial resistance profiles. Blood samples spiked with a clinical isolate of E. coli CCUG 17620 (Culture Collection University of Gothenburg) were also tested to see the effect of elevated incubation temperature on bacterial growth in blood cultures. Results: We observed that increased incubation temperature (42°C) along with vitamin B12 supplementation significantly reduced the tlag (10 – 115 minutes or 4% - 49%) in pure clinical isolates and blood samples spiked with E. coli CCUG17620. In the case of the blood sample, PCR results also detected bacterial DNA after only 3h of incubation and at three times the CFU/mL. Conclusion: Enrichment of bacterial culture media with growth supplements such as vitamin B12 and increased incubation temperature can be a cheap and rapid method for the early detection of pathogens. This proof-of-concept study is restricted to a few bacterial strains and growth conditions. In the future, the effect of other growth conditions and difficult-to-culture bacteria should be explored to shorten the lag phase.

Whole-Genome Sequencing-Based Resistome Analysis of Nosocomial Multidrug-Resistant Non-Fermenting Gram-Negative Pathogens from the Balkans

Non-fermenting Gram-negative bacilli (NFGNB), such as Pseudomonas aeruginosa and Acinetobacter baumannii, are among the major opportunistic pathogens involved in the global antibiotic resistance epidemic. They are designated as urgent/serious threats by the Centers for Disease Control and Prevention and are part of the World Health Organization’s list of critical priority pathogens. Also, Stenotrophomonas maltophilia is increasingly recognized as an emerging cause for healthcare-associated infections in intensive care units, life-threatening diseases in immunocompromised patients, and severe pulmonary infections in cystic fibrosis and COVID-19 individuals. The last annual report of the ECDC showed drastic differences in the proportions of NFGNB with resistance towards key antibiotics in different European Union/European Economic Area countries. The data for the Balkans are of particular concern, indicating more than 80% and 30% of invasive Acinetobacter spp. and P. aeruginosa isolates, respectively, to be carbapenem-resistant. Moreover, multidrug-resistant and extensively drug-resistant S. maltophilia from the region have been recently reported. The current situation in the Balkans includes a migrant crisis and reshaping of the Schengen Area border. This results in collision of diverse human populations subjected to different protocols for antimicrobial stewardship and infection control. The present review article summarizes the findings of whole-genome sequencing-based resistome analyses of nosocomial multidrug-resistant NFGNBs in the Balkan countries.

Increased growth temperature and vitamin B12 supplementation reduces the lag time for rapid pathogen identification in BHI agar and blood cultures

Background: The rapid diagnostics of pathogens is essential to prescribe appropriate and early antibiotic therapy. The current methods for pathogen detection require the bacteria to grow in a culture medium, which is time-consuming. This increases the mortality rate and the global burden of antimicrobial resistance. Culture-free detection methods are still under development and are not used in the clinical routine. Therefore decreasing the culture time for accurate detection of infection and resistance is vital for diagnosis. Methods: In this study, we wanted to investigate easy-to-implement factors (in a minimal laboratory set-up), including inoculum size, incubation temperature, and additional supplementation (e.g., vitamin B12 and trace metals), that can significantly reduce the lag time (tlag). These factors were arranged in simple two-level factorial designs using Gram-positive (Escherichia coli and Pseudomonas aeruginosa) and Gram-negative (Staphylococcus aureus and Bacillus subtilis) bacteria, including clinical isolates with known antimicrobial resistance profiles. Blood samples spiked with a clinical isolate of E. coli CCUG17620 were also tested to see the effect of elevated incubation temperature on bacterial growth in blood cultures. Results: We observed that increased incubation temperature (42°C) along with vitamin B12 supplementation significantly reduced the tlag (10 – 115 minutes or 4% - 49%) in pure clinical isolates and blood samples spiked with E. coli CCUG17620. In the case of the blood sample, PCR results also detected bacterial DNA after only 3h of incubation and at three times the CFU/mL. Conclusions: Enrichment of bacterial culture media with growth supplements such as vitamin B12 and increased incubation temperature can be a cheap and rapid method for the early detection of pathogens. This is a proof-of-concept study restricted to a few bacterial strains and growth conditions. In the future, the effect of other growth conditions and difficult-to-culture bacteria should be explored to shorten the lag phase.

Dimensionality reduction and ensemble of LSTMs for antimicrobial resistance prediction

Bacterial resistance to antibiotics has been rapidly increasing, resulting in low antibiotic effectiveness even treating common infections. The presence of resistant pathogens in environments such as a hospital Intensive Care Unit (ICU) exacerbates the critical admission-acquired infections. This work focuses on the prediction of antibiotic resistance in Pseudomonas aeruginosa nosocomial infections at the ICU, using Long Short-Term Memory (LSTM) artificial neural networks as the predictive method. The analyzed data were extracted from the Electronic Health Records (EHR) of patients admitted to the University Hospital of Fuenlabrada from 2004 to 2019 and were modeled as Multivariate Time Series. A data-driven dimensionality reduction method is built by adapting three feature importance techniques from the literature to the considered data and proposing an algorithm for selecting the most appropriate number of features. This is done using LSTM sequential capabilities so that the temporal aspect of features is taken into account. Furthermore, an ensemble of LSTMs is used to reduce the variance in performance. Our results indicate that the patient's admission information, the antibiotics administered during the ICU stay, and the previous antimicrobial resistance are the most important risk factors. Compared to other conventional dimensionality reduction schemes, our approach is able to improve performance while reducing the number of features for most of the experiments. In essence, the proposed framework achieve, in a computationally cost-efficient manner, promising results for supporting decisions in this clinical task, characterized by high dimensionality, data scarcity, and concept drift.

Application of metagenomic next-generation sequencing in the diagnosis of urinary tract infection in patients undergoing cutaneous ureterostomy

Objective

Urinary tract infection (UTI) is an inflammatory response of the urothelium to bacterial invasion and is a common complication in patients with cutaneous ureterostomy (CU). For such patients, accurate and efficient identification of pathogens remains a challenge. The aim of this study included exploring utility of metagenomic next-generation sequencing (mNGS) in assisting microbiological diagnosis of UTI among patients undergoing CU, identifying promising cytokine or microorganism biomarkers, revealing microbiome diversity change and compare virulence factors (VFs) and antibiotic resistance genes (ARGs) after infection.

Methods

We performed a case-control study of 50 consecutive CU patients from December 2020 to January 2021. According to the clinical diagnostic criteria, samples were divided into infected group and uninfected group and difference of urine culture, cytokines, microorganism, ARGs and VFs were compared between the two groups.

Results

Inflammatory responses were more serious in infected group, as evidenced by a significant increase in IFN-α (p=0.031), IL-1β (0.023) and IL-6 (p=0.018). Clinical culture shows that there is higher positive rate in infected group for most clinical pathogens like Escherichia coli, Klebsiella pneumoniae, Staphylococcus aureus, Candida auris etc. and the top three pathogens with positive frequencies were E. coli, K. pneumoniae, and Enterococcus faecalis. Benchmarking clinical culture, the total sensitivity is 91.4% and specificity is 76.3% for mNGS. As for mNGS, there was no significant difference in microbiome α- diversity between infected and uninfected group. Three species biomarkers including Citrobacter freundii, Klebsiella oxytoca, and Enterobacter cloacae are enriched in infected group based on Lefse. E. cloacae were significantly correlated with IL-6 and IL-10. K. oxytoca were significantly correlated with IL-1β. Besides, the unweighted gene number and weighted gene abundance of VFs or ARGs are significantly higher in infected group. Notablely, ARGs belonging to fluoroquinolones, betalatmas, fosfomycin, phenicol, phenolic compound abundance is significantly higher in infected group which may have bad effect on clinical treatment for patients.

Conclusion

mNGS, along with urine culture, will provide comprehensive and efficient reference for the diagnosis of UTI in patients with CU and allow us to monitor microbial changes in urine of these patients. Moreover, cytokines (IL-6, IL-1β, and IFN-a) or microorganisms like C. freundii, K. oxytoca or E. cloacae are promising biomarkers for building effective UTI diagnostic model of patients with CU and seriously the VFs and ARGs abundance increase in infected group may play bad effect on clinical treatment.

Ceftazidime resistance in Pseudomonas aeruginosa is multigenic and complex

Pseudomonas aeruginosa causes a wide range of severe infections. Ceftazidime, a cephalosporin, is a key antibiotic for treating infections but a significant proportion of isolates are ceftazidime-resistant. The aim of this research was to identify mutations that contribute to resistance, and to quantify the impacts of individual mutations and mutation combinations. Thirty-five mutants with reduced susceptibility to ceftazidime were evolved from two antibiotic-sensitive P. aeruginosa reference strains PAO1 and PA14. Mutations were identified by whole genome sequencing. The evolved mutants tolerated ceftazidime at concentrations between 4 and 1000 times that of the parental bacteria, with most mutants being ceftazidime resistant (minimum inhibitory concentration [MIC] ≥ 32 mg/L). Many mutants were also resistant to meropenem, a carbapenem antibiotic. Twenty-eight genes were mutated in multiple mutants, with dacB and mpl being the most frequently mutated. Mutations in six key genes were engineered into the genome of strain PAO1 individually and in combinations. A dacB mutation by itself increased the ceftazidime MIC by 16-fold although the mutant bacteria remained ceftazidime sensitive (MIC < 32 mg/L). Mutations in ampC, mexR, nalC or nalD increased the MIC by 2- to 4-fold. The MIC of a dacB mutant was increased when combined with a mutation in ampC, rendering the bacteria resistant, whereas other mutation combinations did not increase the MIC above those of single mutants. To determine the clinical relevance of mutations identified through experimental evolution, 173 ceftazidime-resistant and 166 sensitive clinical isolates were analysed for the presence of sequence variants that likely alter function of resistance-associated genes. dacB and ampC sequence variants occur most frequently in both resistant and sensitive clinical isolates. Our findings quantify the individual and combinatorial effects of mutations in different genes on ceftazidime susceptibility and demonstrate that the genetic basis of ceftazidime resistance is complex and multifactorial.

Distinct Long- and Short-Term Adaptive Mechanisms in Pseudomonas aeruginosa

Heterogeneous environments such as the chronically infected cystic fibrosis lung drive the diversification of Pseudomonas aeruginosa populations into, e.g., mucoid, alginate-overproducing isolates or small-colony variants (SCVs). In this study, we performed extensive genome and transcriptome profiling on a clinical SCV isolate that exhibited high cyclic diguanylate (c-di-GMP) levels and a mucoid phenotype. We observed a delayed, stepwise decrease of the high levels of c-di-GMP as well as alginate gene expression upon passaging the SCV under noninducing, rich medium growth conditions over 7 days. Upon prolonged passaging, this lagging reduction of the high c-di-GMP levels under noninducing planktonic conditions (reminiscent of a hysteretic response) was followed by a phenotypic switch to a large-colony morphology, which could be linked to mutations in the Gac/Rsm signaling pathway. Complementation of the Gac/Rsm signaling-negative large-colony variants with a functional GacSA system restored the SCV colony morphotype but was not able to restore the high c-di-GMP levels of the SCV. Our data thus suggest that expression of the SCV colony morphotype and modulation of c-di-GMP levels are genetically separable and follow different evolutionary paths. The delayed switching of c-di-GMP levels in response to fluctuating environmental conditions might provide a unique opportunity to include a time dimension to close the gap between short-term phenotypic and long-term genetic adaptation to biofilm-associated growth conditions. IMPORTANCE Extreme environments, such as those encountered during an infection process in the human host, make effective bacterial adaptation inevitable. While bacteria adapt individually by activating stress responses, long-term adaptation of bacterial communities to challenging conditions can be achieved via genetic fixation of favorable traits. In this study, we describe a two-pronged bacterial stress resistance strategy in the opportunistic pathogen Pseudomonas aeruginosa. We show that the production of adjusted elevated c-di-GMP levels, which drive protected biofilm-associated phenotypes in vivo, resembles a stable hysteretic response which prevents unwanted frequent switching. Cellular hysteresis might provide a link between individual adaptability and evolutionary adaptation to ensure the evolutionary persistence of host-adapted stress response strategies.

Core Genome Multilocus Sequence Typing and Antibiotic Susceptibility Prediction from Whole-Genome Sequence Data of Multidrug-Resistant Pseudomonas aeruginosa Isolates

Over the past decade, whole-genome sequencing (WGS) has overtaken traditional bacterial typing methods for studies of genetic relatedness. Further, WGS data generated during epidemiologic studies can be used in other clinically relevant bioinformatic applications, such as antibiotic resistance prediction. Using commercially available software tools, the relatedness of 38 clinical isolates of multidrug-resistant Pseudomonas aeruginosa was defined by two core genome multilocus sequence typing (cgMLST) methods, and the WGS data of each isolate was analyzed to predict antibiotic susceptibility to nine antibacterial agents. The WGS typing and resistance prediction data were compared with pulsed-field gel electrophoresis (PFGE) and phenotypic antibiotic susceptibility results, respectively. Simpson's Diversity Index and adjusted Wallace pairwise assessments of the three typing methods showed nearly identical discriminatory power. Antibiotic resistance prediction using a trained analytical pipeline examined 342 bacterial-drug combinations with an overall categorical agreement of 92.4% and very major, major, and minor error rates of 3.6, 4.1, and 4.1%, respectively. IMPORTANCE Multidrug-resistant Pseudomonas aeruginosa isolates are a serious public health concern due to their resistance to nearly all or all of the available antibiotics, including carbapenems. Utilizing molecular approaches in conjunction with antibiotic susceptibility prediction software warrants investigation for use in the clinical laboratory workflow. These molecular tools coupled with antibiotic resistance prediction tools offer the opportunity to overcome the extended turnaround time and technical challenges of phenotypic susceptibility testing.

Application of Decision-Tree-Based Machine Learning Algorithms for Prediction of Antimicrobial Resistance

Timely and efficacious antibiotic treatment depends on precise and quick in silico antimicrobial-resistance predictions. Limited treatment choices due to antimicrobial resistance (AMR) highlight the necessity to optimize the available diagnostics. AMR can be explicitly anticipated on the basis of genome sequence. In this study, we used transcriptomes of 410 multidrug-resistant isolates of Pseudomonas aeruginosa. We trained 10 machine learning (ML) classifiers on the basis of data on gene expression (GEXP) information and generated predictive models for meropenem, ciprofloxacin, and ceftazidime drugs. Among all the used ML models, four models showed high F1-score, accuracy, precision, and specificity compared with the other models. However, RandomForestClassifier showed a moderate F1-score (0.6), precision (0.61), and specificity (0.625) for ciprofloxacin. In the case of ceftazidime, RidgeClassifier performed well and showed F1-score (0.652), precision (0.654), and specificity (0.652) values. For meropenem, KNeighborsClassifier exhibited moderate F1-score (0.629), precision (0.629), and specificity (0.629). Among these three antibiotics, GEXP data on meropenem and ceftazidime improved diagnostic performance. The findings will pave the way for the establishment of a resistance profiling tool that can predict AMR on the basis of transcriptomic markers.

Pseudomonas aeruginosa Production of Hydrogen Cyanide Leads to Airborne Control of Staphylococcus aureus Growth in Biofilm and In Vivo Lung Environments

Diverse bacterial volatile compounds alter bacterial stress responses and physiology, but their contribution to population dynamics in polymicrobial communities is not well known. In this study, we showed that airborne volatile hydrogen cyanide (HCN) produced by a wide range of Pseudomonas aeruginosa clinical strains leads to at-a-distance in vitro inhibition of the growth of a wide array of Staphylococcus aureus strains. We determined that low-oxygen environments not only enhance P. aeruginosa HCN production but also increase S. aureus sensitivity to HCN, which impacts P. aeruginosa-S. aureus competition in microaerobic in vitro mixed biofilms as well as in an in vitro cystic fibrosis lung sputum medium. Consistently, we demonstrated that production of HCN by P. aeruginosa controls S. aureus growth in a mouse model of airways coinfected by P. aeruginosa and S. aureus. Our study therefore demonstrates that P. aeruginosa HCN contributes to local and distant airborne competition against S. aureus and potentially other HCN-sensitive bacteria in contexts relevant to cystic fibrosis and other polymicrobial infectious diseases. IMPORTANCE Airborne volatile compounds produced by bacteria are often only considered attractive or repulsive scents, but they also directly contribute to bacterial physiology. Here, we showed that volatile hydrogen cyanide (HCN) released by a wide range of Pseudomonas aeruginosa strains controls Staphylococcus aureus growth in low-oxygen in vitro biofilms or aggregates and in vivo lung environments. These results are of pathophysiological relevance, since lungs of cystic fibrosis patients are known to present microaerobic areas and to be commonly associated with the presence of S. aureus and P. aeruginosa in polymicrobial communities. Our study therefore provides insights into how a bacterial volatile compound can contribute to the exclusion of S. aureus and other HCN-sensitive competitors from P. aeruginosa ecological niches. It opens new perspectives for the management or monitoring of P. aeruginosa infections in lower-lung airway infections and other polymicrobial disease contexts.

Resistance to Fluoroquinolones in Pseudomonas aeruginosa from Human, Animal, Food and Environmental Origin: The Role of CrpP and Mobilizable ICEs

Fluoroquinolone resistance and the associated genetic mechanisms were assessed by antimicrobial susceptibility and whole genome sequencing in 56 Pseudomonas aeruginosa strains from human, animal, food and environmental origins. P. aeruginosa PAO1, PA7 and PA14 reference strains were also included in the study. Twenty-two strains (37%) were resistant to, at least, one fluoroquinolone agent. Correlation between the number of changes in GyrA and ParC proteins and the level of fluoroquinolone resistance was observed. Mutations or absence of genes, such as mexZ, mvaT and nalD encoding efflux pumps regulators, were also found in resistant strains. The crpP gene was detected in 43 strains (72.9%; 17 of them non-clinical strains), and coded seven different CrpP variants, including a novel one (CrpP-7). The crpP gene was located in 23 different chromosomal mobile integrative and conjugative elements (ICEs), inserted in two tRNAs integration sites. A great variety of structures was detected in the crpP-ICEs elements, e.g., the fimbriae related cup clusters, the mercury resistance mer operon, the pyocin S5 or S8 bacteriocin encoding genes, and mobilization genes. The location of crpP-like genes in mobilizable ICEs and linked to heavy metal resistance and virulence factors is of significant concern in P. aeruginosa. This work provides a genetic explanation of the fluoroquinolone resistance and crpP-associated pathogenesis of P. aeruginosa from a One-Health approach.

Paving the way for precise diagnostics of antimicrobial resistant bacteria

The antimicrobial resistance (AMR) crisis from bacterial pathogens is frequently emerging and rapidly disseminated during the sustained antimicrobial exposure in human-dominated communities, posing a compelling threat as one of the biggest challenges in humans. The frequent incidences of some common but untreatable infections unfold the public health catastrophe that antimicrobial-resistant pathogens have outpaced the available countermeasures, now explicitly amplified during the COVID-19 pandemic. Nowadays, biotechnology and machine learning advancements help create more fundamental knowledge of distinct spatiotemporal dynamics in AMR bacterial adaptation and evolutionary processes. Integrated with reliable diagnostic tools and powerful analytic approaches, a collaborative and systematic surveillance platform with high accuracy and predictability should be established and implemented, which is not just for an effective controlling strategy on AMR but also for protecting the longevity of valuable antimicrobials currently and in the future.

Effects of the Sex Steroid Hormone Estradiol on Biofilm Growth of Cystic Fibrosis Pseudomonas aeruginosa Isolates

Women with cystic fibrosis (CF) have a significantly lower life expectancy compared to men, which is indicated by an earlier impairment of lung function due to chronic colonization with biofilm formed by Pseudomonas aeruginosa. There is growing evidence that blood serum concentrations of the steroid sex hormone estradiol (E₂) correlate with the occurrence of pulmonary exacerbations in CF but also play a role in the mucoid switch of P. aeruginosa. This study aims to shed light on possible microbiological reasons for sexual dimorphism in CF by investigating the influence of E₂ on biofilm formation of P. aeruginosa CF isolates. For this purpose, 10 CF isolates of the respiratory tract derived from different CF patients have been treated with E₂ in a microtiter plate biofilm model. Biofilms have been examined by crystal violet assays, field emission scanning electron microscopy (FE-SEM), 3D laser scanning microscopy (LSM), and quorum sensing (QS) reporter assays of the supernatants taken from biofilms. This allowed us to simultaneously investigate the effects of E₂ on attached biofilm mass, biofilm ultrastructure, and QS activity. Upon E₂ treatment, six out of 10 investigated CF isolates showed an increase of attached biofilm mass, whereas biofilms from two tested non-CF laboratory strains (PAO1 and ATCC19660) did not. Moreover, FE-SEM and 3D LSM analyses of the E₂ responsive CF biofilms revealed ultrastructural remodeling of biofilm structure at different scales with increased formation of prominent biofilm spots, enhanced coverage with extracellular polymeric substance (EPS), and extended average surface roughness. QS activity measurements performed in biofilm supernatants via luminescence acyl homoserine lactone (AHL) reporter assays further showed that E₂ treatment may also modulate QS signaling, as shown in an E₂ sensitive CF isolate. Together, our results suggest the biofilm modulating effects of E₂ on various clinical CF isolates that are documented by both biomass and ultrastructural changes of biofilms. The gained new insight into the influence of steroid hormones on P. aeruginosa biofilm phenotypes might pave the way for novel future approaches in personalized medicine based on the patients’ sex and hormonal status.

Antibiotic resistance in chronic respiratory diseases: from susceptibility testing to the resistome

The development of resistome analysis, i.e. the comprehensive analysis of antibiotic-resistance genes (ARGs), is enabling a better understanding of the mechanisms of antibiotic-resistance emergence. The respiratory microbiome is a dynamic and interactive network of bacteria, with a set of ARGs that could influence the response to antibiotics. Viruses such as bacteriophages, potential carriers of ARGs, may also form part of this respiratory resistome. Chronic respiratory diseases (CRDs) such as cystic fibrosis, severe asthma, chronic obstructive pulmonary disease and bronchiectasis, managed with long-term antibiotic therapies, lead to multidrug resistance. Antibiotic susceptibility testing provides a partial view of the bacterial response to antibiotics in the complex lung environment. Assessing the ARG network would allow personalised, targeted therapeutic strategies and suitable antibiotic stewardship in CRDs, depending on individual resistome and microbiome signatures. This review summarises the influence of pulmonary antibiotic protocols on the respiratory microbiome, detailing the variable consequences according to antibiotic class and duration of treatment. The different resistome-profiling methods are explained to clarify their respective place in antibiotic-resistance analysis in the lungs. Finally, this review details current knowledge on the respiratory resistome related to therapeutic strategies and provides insight into the application of resistome analysis to counter the emergence of multidrug-resistant respiratory pathogens.

Application of Artificial Intelligence in Combating High Antimicrobial Resistance Rates

Artificial intelligence (AI) is a branch of science and engineering that focuses on the computational understanding of intelligent behavior. Many human professions, including clinical diagnosis and prognosis, are greatly useful from AI. Antimicrobial resistance (AMR) is among the most critical challenges facing Pakistan and the rest of the world. The rising incidence of AMR has become a significant issue, and authorities must take measures to combat the overuse and incorrect use of antibiotics in order to combat rising resistance rates. The widespread use of antibiotics in clinical practice has not only resulted in drug resistance but has also increased the threat of super-resistant bacteria emergence. As AMR rises, clinicians find it more difficult to treat many bacterial infections in a timely manner, and therapy becomes prohibitively costly for patients. To combat the rise in AMR rates, it is critical to implement an institutional antibiotic stewardship program that monitors correct antibiotic use, controls antibiotics, and generates antibiograms. Furthermore, these types of tools may aid in the treatment of patients in the event of a medical emergency in which a physician is unable to wait for bacterial culture results. AI’s applications in healthcare might be unlimited, reducing the time it takes to discover new antimicrobial drugs, improving diagnostic and treatment accuracy, and lowering expenses at the same time. The majority of suggested AI solutions for AMR are meant to supplement rather than replace a doctor’s prescription or opinion, but rather to serve as a valuable tool for making their work easier. When it comes to infectious diseases, AI has the potential to be a game-changer in the battle against antibiotic resistance. Finally, when selecting antibiotic therapy for infections, data from local antibiotic stewardship programs are critical to ensuring that these bacteria are treated quickly and effectively. Furthermore, organizations such as the World Health Organization (WHO) have underlined the necessity of selecting the appropriate antibiotic and treating for the shortest time feasible to minimize the spread of resistant and invasive resistant bacterial strains.

Ensemble of Artificial Intelligence Techniques for Bacterial Antimicrobial Resistance (AMR) Estimation Using Topic Modeling and Similarity Measure

In recent times, bacterial Antimicrobial Resistance (AMR) analyses becomes a hot study topic. The AMR comprises information related to the antibiotic product name, class name, subclass name, type, subtype, gene type, etc., which can fight against the illness. However, the tagging language used to determine the data is of free context. These contexts often contain ambiguous data, which leads to a hugely challenging issue in retrieving, organizing, merging, and finding the relevant data. Manually reading this text and labelling is not time-consuming. Hence, topic modeling overcomes these challenges and provides efficient results in categorizing the topic and in determining the data. In this view, this research work designs an ensemble of artificial intelligence for categorizing the AMR gene data and determine the relationship between the antibiotics. The proposed model includes a weighted voting based ensemble model by the incorporation of Latent Dirichlet Allocation (LDA) and Hierarchical Recurrent Neural Networks (HRNN), shows the novelty of the work. It is used for determining the amount of “topics” that cluster utilizing a multidimensional scaling approach. In addition, the proposed model involves the data pre-processing stage to get rid of stop words, punctuations, lower casing, etc. Moreover, an explanatory data analysis uses word cloud which assures the proper functionality and to proceed with the model training process. Besides, three approaches namely perplexity, Harmonic mean, and Random initialization of K are employed to determine the number of topics. For experimental validation, an openly accessible Bacterial AMR reference gene database is employed. The experimental results reported that the perplexity provided the optimal number of topics from the AMR gene data of more than 6500 samples. Therefore, the proposed model helps to find the appropriate antibiotic for bacterial and viral spread and discover how to increase the proper antibiotic in human bodies

Expanding the search for small-molecule antibacterials by multidimensional profiling

New techniques for systematic profiling of small-molecule effects can enhance traditional growth inhibition screens for antibiotic discovery and change how we search for new antibacterial agents. Computational models that integrate physicochemical compound properties with their phenotypic and molecular downstream effects can not only predict efficacy of molecules yet to be tested, but also reveal unprecedented insights on compound modes of action (MoAs). The unbiased characterization of compounds that themselves are not growth inhibitory but exhibit diverse MoAs, can expand antibacterial strategies beyond direct inhibition of core essential functions. Early and systematic functional annotation of compound libraries thus paves the way to new models in the selection of lead antimicrobial compounds. In this Review, we discuss how multidimensional small-molecule profiling and the ever-increasing computing power are accelerating the discovery of unconventional antibacterials capable of bypassing resistance and exploiting synergies with established antibacterial treatments and with protective host mechanisms.

Inferring antibiotic susceptibility from metagenomic data: dream or reality?

BACKGROUND

The diagnosis of bacterial infections continues to rely on culture, a slow process in which antibiotic susceptibility profiles of potential pathogens are made available to clinicians 48h after sampling, at best. Recently, clinical metagenomics (CMg), the metagenomic sequencing of samples with the purpose of identifying microorganisms and determining their susceptibility to antimicrobials, has emerged as a potential diagnostic tool that could prove faster than culture. CMg indeed has the potential to detect antibiotic resistance genes (ARGs) and mutations associated with resistance. Nevertheless, many challenges have yet to be overcome in order to make rapid phenotypic inference of antibiotic susceptibility from metagenomic data a reality.

OBJECTIVES

The objective of this narrative review is to discuss the challenges underlying the phenotypic inference of antibiotic susceptibility from metagenomic data.

SOURCES

We conducted a narrative review using published articles available in the NCBI Pubmed database.

CONTENT

We review the current ARG databases with a specific emphasis on those which now provide associations with phenotypic data. Next, we discuss the bioinformatic tools designed to identify ARGs in metagenomes. We then report on the performance of phenotypic inference from genomic data and the issue predicting the expression of ARGs. Finally, we address the challenge of linking an ARG to this host.

IMPLICATIONS

Significant improvements have recently been made in associating ARG and phenotype, and the inference of susceptibility from genomic data has been demonstrated in pathogenic bacteria such as Staphylococci and Enterobacterales. Resistance involving gene expression is more challenging however, and inferring susceptibility from species such as Pseudomonas aeruginosa remains difficult. Future research directions include the consideration of gene expression via RNA sequencing and machine learning.

Antibiotic stewardship in the era of precision medicine

Antimicrobial resistance (AMR) continues to spread at an alarming rate worldwide. Novel approaches are needed to mitigate its deleterious impact on antibiotic efficacy. Antibiotic stewardship aims to promote the appropriate use of antibiotics through evidence-based interventions. One paradigm is precision medicine, a medical model in which decisions, practices, interventions, and therapies are adapted to the individual patient based on their predicted response or risk of disease. Precision medicine approaches hold promise as a way to improve outcomes for patients with myriad illnesses, including infections such as bacteraemia and pneumonia. This review describes the latest advances in precision medicine as they pertain to antibiotic stewardship, with an emphasis on hospital-based antibiotic stewardship programmes. The impact of the COVID-19 pandemic on AMR and antibiotic stewardship, gaps in the scientific evidence, and areas for further research are also discussed.

A Practical Approach for Predicting Antimicrobial Phenotype Resistance in Staphylococcus aureus Through Machine Learning Analysis of Genome Data

With the reduction in sequencing price and acceleration of sequencing speed, it is particularly important to directly link the genotype and phenotype of bacteria. Here, we firstly predicted the minimum inhibitory concentrations of ten antimicrobial agents for Staphylococcus aureus using 466 isolates by directly extracting k-mer from whole genome sequencing data combined with three machine learning algorithms: random forest, support vector machine, and XGBoost. Considering one two-fold dilution, the essential agreement and the category agreement could reach >85% and >90% for most antimicrobial agents. For clindamycin, cefoxitin and trimethoprim-sulfamethoxazole, the essential agreement and the category agreement could reach >91% and >93%, providing important information for clinical treatment. The successful prediction of cefoxitin resistance showed that the model could identify methicillin-resistant S. aureus. The results suggest that small datasets available in large hospitals could bypass the existing basic research and known antimicrobial resistance genes and accurately predict the bacterial phenotype.

Persistent Bacterial Infections, Antibiotic Treatment Failure, and Microbial Adaptive Evolution

Antibiotic resistance is expected by the WHO to be the biggest threat to human health before 2050. In this overview, we argue that this prediction may in fact be too optimistic because it is often overlooked that many bacterial infections frequently ‘go under the radar’ because they are difficult to diagnose and characterize. Due to our lifestyle, persistent infections caused by opportunistic bacteria—well-known or emerging—show increasing success of infecting patients with reduced defense capacity, and often antibiotics fail to be sufficiently effective, even if the bacteria are susceptible, leaving small bacterial populations unaffected by treatment in the patient. The mechanisms behind infection persistence are multiple, and therefore very difficult to diagnose in the laboratory and to treat. In contrast to antibiotic resistance associated with acute infections caused by traditional bacterial pathogens, genetic markers associated with many persistent infections are imprecise and mostly without diagnostic value. In the absence of effective eradication strategies, there is a significant risk that persistent infections may eventually become highly resistant to antibiotic treatment due to the accumulation of genomic mutations, which will transform colonization into persistence.

Pseudomonas aeruginosa Is More Tolerant Under Biofilm Than Under Planktonic Growth Conditions: A Multi-Isolate Survey

Biofilm-associated bacteria exhibit profound changes in bacterial physiology. They thrive in the environment but also in the human host in protected sessile communities. Antimicrobial therapy usually fails, despite the absence of genotypic resistance, and it is commonly accepted that biofilm-grown bacteria are up to 1,000-fold more resistant than planktonic cells. We are only at the beginning to understand the reasons for biofilm recalcitrance, and systematic approaches to describe biofilm-induced tolerance phenotypes are lacking. In this study, we investigated a large and highly diverse collection of 352 clinical Pseudomonas aeruginosa isolates for their antimicrobial susceptibility profiles under biofilm growth conditions towards the antibiotics ciprofloxacin, tobramycin, and colistin. We discovered characteristic patterns of drug-specific killing activity and detected conditional tolerance levels far lower (in the range of the minimal inhibitory concentration (MIC)), but also far higher (up to 16,000-fold increase compared to planktonic cells) than generally believed. This extremely broad distribution of biofilm-induced tolerance phenotypes across the clinical isolates was greatly influenced by the choice of the antibiotic. We furthermore describe cross-tolerance against ciprofloxacin and tobramycin, but not colistin, and observed an additive activity between biofilm-induced tolerance and genetically determined resistance. This became less evident when the biofilm-grown cells were exposed to very high antibiotic concentrations. Although much more remains to be learned on the molecular mechanisms underlying biofilm-induced tolerance, our data on intra-species variations in tolerance profiles provide valuable new insights. Furthermore, our observation that colistin appears to act independently of the tolerance mechanisms of individual clinical strains could make colistin a valuable therapeutic option in chronic biofilm-associated infections characterized by the presence of particularly tolerant strains.

Evolution of biofilm-adapted gene expression profiles in lasR-deficient clinical Pseudomonas aeruginosa isolates

The overall success of a pathogenic microbe depends on its ability to efficiently adapt to challenging conditions in the human host. Long-term evolution experiments track and predict adaptive trajectories and have contributed significantly to our understanding of the driving forces of bacterial adaptation. In this study, we conducted a cross-sectional study instead of long-term longitudinal evolution experiments. We analyzed the transcriptional profiles as well as genomic sequence variations of a large number of clinical Pseudomonas aeruginosa isolates that have been recovered from different infected human sites. Convergent changes in gene expression patterns were found in different groups of clinical isolates. The majority of repeatedly observed expression patterns could be attributed to a defective lasR gene, which encodes the major quorum-sensing regulator LasR. Strikingly, the gene expression pattern of the lasR-defective strains appeared to reflect a transcriptional response that evolves in a direction consistent with growth within a biofilm. In a process of genetic assimilation, lasR-deficient P. aeruginosa isolates appear to constitutively express a biofilm-adapted transcriptional profile and no longer require a respective environmental trigger. Our results demonstrate that profiling the functional consequences of pathoadaptive mutations in clinical isolates reveals long-term evolutionary pathways and may explain the success of lasR mutants in the opportunistic pathogen P. aeruginosa in a clinical context.