Convergence of resistance and evolutionary responses in Escherichia coli and Salmonella enterica co-inhabiting chicken farms in China.

Sharing of genetic elements among different pathogens and commensals inhabiting same hosts and environments has significant implications for antimicrobial resistance (AMR), especially in settings with high antimicrobial exposure. We analysed 661 Escherichia coli and Salmonella enterica isolates collected within and across hosts and environments, in 10 Chinese chicken farms over 2.5 years using data-mining methods. Most isolates within same hosts possessed the same clinically relevant AMR-carrying mobile genetic elements (plasmids: 70.6%, transposons: 78%), which also showed recent common evolution. Supervised machine learning classifiers revealed known and novel AMR-associated mutations and genes underlying resistance to 28 antimicrobials, primarily associated with resistance in E. coli and susceptibility in S. enterica. Many were essential and affected same metabolic processes in both species, albeit with varying degrees of phylogenetic penetration. Multi-modal strategies are crucial to investigate the interplay of mobilome, resistance and metabolism in cohabiting bacteria, especially in ecological settings where community-driven resistance selection occurs.

Machine learning and metagenomics reveal shared antimicrobial resistance profiles across multiple chicken farms and abattoirs in China.

China is the largest global consumer of antimicrobials and improving surveillance methods could help to reduce antimicrobial resistance (AMR) spread. Here we report the surveillance of ten large-scale chicken farms and four connected abattoirs in three Chinese provinces over 2.5 years. Using a data mining approach based on machine learning, we analysed 461 microbiomes from birds, carcasses and environments, identifying 145 potentially mobile antibiotic resistance genes (ARGs) shared between chickens and environments across all farms. A core set of 233 ARGs and 186 microbial species extracted from the chicken gut microbiome correlated with the AMR profiles of Escherichia coli colonizing the same gut, including Arcobacter, Acinetobacter and Sphingobacterium, clinically relevant for humans, and 38 clinically relevant ARGs. Temperature and humidity in the barns were also correlated with ARG presence. We reveal an intricate network of correlations between environments, microbial communities and AMR, suggesting multiple routes to improving AMR surveillance in livestock production.

Activated tissue resident memory T-cells (CD8+CD103+CD39+) uniquely predict survival in left sided "immune-hot" colorectal cancers.

Introduction: Characterization of the tumour immune infiltrate (notably CD8+ T-cells) has strong predictive survival value for cancer patients. Quantification of CD8 T-cells alone cannot determine antigenic experience, as not all infiltrating T-cells recognize tumour antigens. Activated tumour-specific tissue resident memory CD8 T-cells (TRM) can be defined by the co-express of CD103, CD39 and CD8. We investigated the hypothesis that the abundance and localization of TRM provides a higher-resolution route to patient stratification. Methods: A comprehensive series of 1000 colorectal cancer (CRC) were arrayed on a tissue microarray, with representative cores from three tumour locations and the adjacent normal mucosa. Using multiplex immunohistochemistry we quantified and determined the localization of TRM. Results: Across all patients, activated TRM were an independent predictor of survival, and superior to CD8 alone. Patients with the best survival had immune-hot tumours heavily infiltrated throughout with activated TRM. Interestingly, differences between right- and left-sided tumours were apparent. In left-sided CRC, only the presence of activated TRM (and not CD8 alone) was prognostically significant. Patients with low numbers of activated TRM cells had a poor prognosis even with high CD8 T-cell infiltration. In contrast, in right-sided CRC, high CD8 T-cell infiltration with low numbers of activated TRM was a good prognosis. Conclusion: The presence of high intra-tumoural CD8 T-cells alone is not a predictor of survival in left-sided CRC and potentially risks under treatment of patients. Measuring both high tumour-associated TRM and total CD8 T-cells in left-sided disease has the potential to minimize current under-treatment of patients. The challenge will be to design immunotherapies, for left-sided CRC patients with high CD8 T-cells and low activate TRM,that result in effective immune responses and thereby improve patient survival.

Dissecting microbial communities and resistomes for interconnected humans, soil, and livestock.

A debate is currently ongoing as to whether intensive livestock farms may constitute reservoirs of clinically relevant antimicrobial resistance (AMR), thus posing a threat to surrounding communities. Here, combining shotgun metagenome sequencing, machine learning (ML), and culture-based methods, we focused on a poultry farm and connected slaughterhouse in China, investigating the gut microbiome of livestock, workers and their households, and microbial communities in carcasses and soil. For both the microbiome and resistomes in this study, differences are observed across environments and hosts. However, at a finer scale, several similar clinically relevant antimicrobial resistance genes (ARGs) and similar associated mobile genetic elements were found in both human and broiler chicken samples. Next, we focused on Escherichia coli, an important indicator for the surveillance of AMR on the farm. Strains of E. coli were found intermixed between humans and chickens. We observed that several ARGs present in the chicken faecal resistome showed correlation to resistance/susceptibility profiles of E. coli isolates cultured from the same samples. Finally, by using environmental sensing these ARGs were found to be correlated to variations in environmental temperature and humidity. Our results show the importance of adopting a multi-domain and multi-scale approach when studying microbial communities and AMR in complex, interconnected environments.

Whole-genome sequencing and gene sharing network analysis powered by machine learning identifies antibiotic resistance sharing between animals, humans and environment in livestock farming.

Anthropogenic environments such as those created by intensive farming of livestock, have been proposed to provide ideal selection pressure for the emergence of antimicrobial-resistant Escherichia coli bacteria and antimicrobial resistance genes (ARGs) and spread to humans. Here, we performed a longitudinal study in a large-scale commercial poultry farm in China, collecting E. coli isolates from both farm and slaughterhouse; targeting animals, carcasses, workers and their households and environment. By using whole-genome phylogenetic analysis and network analysis based on single nucleotide polymorphisms (SNPs), we found highly interrelated non-pathogenic and pathogenic E. coli strains with phylogenetic intermixing, and a high prevalence of shared multidrug resistance profiles amongst livestock, human and environment. Through an original data processing pipeline which combines omics, machine learning, gene sharing network and mobile genetic elements analysis, we investigated the resistance to 26 different antimicrobials and identified 361 genes associated to antimicrobial resistance (AMR) phenotypes; 58 of these were known AMR-associated genes and 35 were associated to multidrug resistance. We uncovered an extensive network of genes, correlated to AMR phenotypes, shared among livestock, humans, farm and slaughterhouse environments. We also found several human, livestock and environmental isolates sharing closely related mobile genetic elements carrying ARGs across host species and environments. In a scenario where no consensus exists on how antibiotic use in the livestock may affect antibiotic resistance in the human population, our findings provide novel insights into the broader epidemiology of antimicrobial resistance in livestock farming. Moreover, our original data analysis method has the potential to uncover AMR transmission pathways when applied to the study of other pathogens active in other anthropogenic environments characterised by complex interconnections between host species.

Genome-Scale Metabolic Models and Machine Learning Reveal Genetic Determinants of Antibiotic Resistance in Escherichia coli and Unravel the Underlying Metabolic Adaptation Mechanisms.

Antimicrobial resistance (AMR) is becoming one of the largest threats to public health worldwide, with the opportunistic pathogen Escherichia coli playing a major role in the AMR global health crisis. Unravelling the complex interplay between drug resistance and metabolic rewiring is key to understand the ability of bacteria to adapt to new treatments and to the development of new effective solutions to combat resistant infections. We developed a computational pipeline that combines machine learning with genome-scale metabolic models (GSMs) to elucidate the systemic relationships between genetic determinants of resistance and metabolism beyond annotated drug resistance genes. Our approach was used to identify genetic determinants of 12 AMR profiles for the opportunistic pathogenic bacterium E. coli. Then, to interpret the large number of identified genetic determinants, we applied a constraint-based approach using the GSM to predict the effects of genetic changes on growth, metabolite yields, and reaction fluxes. Our computational platform leads to multiple results. First, our approach corroborates 225 known AMR-conferring genes, 35 of which are known for the specific antibiotic. Second, integration with the GSM predicted 20 top-ranked genetic determinants (including accA, metK, fabD, fabG, murG, lptG, mraY, folP, and glmM) essential for growth, while a further 17 top-ranked genetic determinants linked AMR to auxotrophic behavior. Third, clusters of AMR-conferring genes affecting similar metabolic processes are revealed, which strongly suggested that metabolic adaptations in cell wall, energy, iron and nucleotide metabolism are associated with AMR. The computational solution can be used to study other human and animal pathogens. IMPORTANCE Escherichia coli is a major public health concern given its increasing level of antibiotic resistance worldwide and extraordinary capacity to acquire and spread resistance via horizontal gene transfer with surrounding species and via mutations in its existing genome. E. coli also exhibits a large amount of metabolic pathway redundancy, which promotes resistance via metabolic adaptability. In this study, we developed a computational approach that integrates machine learning with metabolic modeling to understand the correlation between AMR and metabolic adaptation mechanisms in this model bacterium. Using our approach, we identified AMR genetic determinants associated with cell wall modifications for increased permeability, virulence factor manipulation of host immunity, reduction of oxidative stress toxicity, and changes to energy metabolism. Unravelling the complex interplay between antibiotic resistance and metabolic rewiring may open new opportunities to understand the ability of E. coli, and potentially of other human and animal pathogens, to adapt to new treatments.

Whole-Genome Sequencing and Machine Learning Analysis of Staphylococcus aureus from Multiple Heterogeneous Sources in China Reveals Common Genetic Traits of Antimicrobial Resistance.

Staphylococcus aureus is a worldwide leading cause of numerous diseases ranging from food-poisoning to lethal infections. Methicillin-resistant S. aureus (MRSA) has been found capable of acquiring resistance to most antimicrobials. MRSA is ubiquitous and diverse even in terms of antimicrobial resistance (AMR) profiles, posing a challenge for treatment. Here, we present a comprehensive study of S. aureus in China, addressing epidemiology, phylogenetic reconstruction, genomic characterization, and identification of AMR profiles. The study analyzes 673 S. aureus isolates from food as well as from hospitalized and healthy individuals. The isolates have been collected over a 9-year period, between 2010 and 2018, from 27 provinces across China. By whole-genome sequencing, Bayesian divergence analysis, and supervised machine learning, we reconstructed the phylogeny of the isolates and compared them to references from other countries. We identified 72 sequence types (STs), of which, 29 were novel. We found 81 MRSA lineages by multilocus sequence type (MLST), spa, staphylococcal cassette chromosome mec element (SCCmec), and Panton-Valentine leukocidin (PVL) typing. In addition, novel variants of SCCmec type IV hosting extra metal and antimicrobial resistance genes, as well as a new SCCmec type, were found. New Bayesian dating of the split times of major clades showed that ST9, ST59, and ST239 in China and European countries fell in different branches, whereas this pattern was not observed for the ST398 clone. On the contrary, the clonal transmission of ST398 was more intermixed in regard to geographic origin. Finally, we identified genetic determinants of resistance to 10 antimicrobials, discriminating drug-resistant bacteria from susceptible strains in the cohort. Our results reveal the emergence of Chinese MRSA lineages enriched of AMR determinants that share similar genetic traits of antimicrobial resistance across human and food, hinting at a complex scenario of evolving transmission routes. IMPORTANCE Little information is available on the epidemiology and characterization of Staphylococcus aureus in China. The role of food is a cause of major concern: staphylococcal foodborne diseases affect thousands every year, and the presence of resistant Staphylococcus strains on raw retail meat products is well documented. We studied a large heterogeneous data set of S. aureus isolates from many provinces of China, isolated from food as well as from individuals. Our large whole-genome collection represents a unique catalogue that can be easily meta-analyzed and integrated with further studies and adds to the library of S. aureus sequences in the public domain in a currently underrepresented geographical region. The new Bayesian dating of the split times of major drug-resistant enriched clones is relevant in showing that Chinese and European methicillin-resistant S. aureus (MRSA) have evolved differently. Our machine learning approach, across a large number of antibiotics, shows novel determinants underlying resistance and reveals frequent resistant traits in specific clonal complexes, highlighting the importance of particular clonal complexes in China. Our findings substantially expand what is known of the evolution and genetic determinants of resistance in food-associated S. aureus in China and add crucial information for whole-genome sequencing (WGS)-based surveillance of S. aureus.

Mass spectrometry and machine learning for the accurate diagnosis of benzylpenicillin and multidrug resistance of Staphylococcus aureus in bovine mastitis.

Staphylococcus aureus is a serious human and animal pathogen threat exhibiting extraordinary capacity for acquiring new antibiotic resistance traits in the pathogen population worldwide. The development of fast, affordable and effective diagnostic solutions capable of discriminating between antibiotic-resistant and susceptible S. aureus strains would be of huge benefit for effective disease detection and treatment. Here we develop a diagnostics solution that uses Matrix-Assisted Laser Desorption/Ionisation-Time of Flight Mass Spectrometry (MALDI-TOF) and machine learning, to identify signature profiles of antibiotic resistance to either multidrug or benzylpenicillin in S. aureus isolates. Using ten different supervised learning techniques, we have analysed a set of 82 S. aureus isolates collected from 67 cows diagnosed with bovine mastitis across 24 farms. For the multidrug phenotyping analysis, LDA, linear SVM, RBF SVM, logistic regression, naïve Bayes, MLP neural network and QDA had Cohen's kappa values over 85.00%. For the benzylpenicillin phenotyping analysis, RBF SVM, MLP neural network, naïve Bayes, logistic regression, linear SVM, QDA, LDA, and random forests had Cohen's kappa values over 85.00%. For the benzylpenicillin the diagnostic systems achieved up to (mean result ± standard deviation over 30 runs on the test set): accuracy = 97.54% ± 1.91%, sensitivity = 99.93% ± 0.25%, specificity = 95.04% ± 3.83%, and Cohen's kappa = 95.04% ± 3.83%. Moreover, the diagnostic platform complemented by a protein-protein network and 3D structural protein information framework allowed the identification of five molecular determinants underlying the susceptible and resistant profiles. Four proteins were able to classify multidrug-resistant and susceptible strains with 96.81% ± 0.43% accuracy. Five proteins, including the previous four, were able to classify benzylpenicillin resistant and susceptible strains with 97.54% ± 1.91% accuracy. Our approach may open up new avenues for the development of a fast, affordable and effective day-to-day diagnostic solution, which would offer new opportunities for targeting resistant bacteria.

Prediction of Streptococcus uberis clinical mastitis treatment success in dairy herds by means of mass spectrometry and machine-learning.

Streptococcus uberis is one of the leading pathogens causing mastitis worldwide. Identification of S. uberis strains that fail to respond to treatment with antibiotics is essential for better decision making and treatment selection. We demonstrate that the combination of supervised machine learning and matrix-assisted laser desorption ionization/time of flight (MALDI-TOF) mass spectrometry can discriminate strains of S. uberis causing clinical mastitis that are likely to be responsive or unresponsive to treatment. Diagnostics prediction systems trained on 90 individuals from 26 different farms achieved up to 86.2% and 71.5% in terms of accuracy and Cohen's kappa. The performance was further increased by adding metadata (parity, somatic cell count of previous lactation and count of positive mastitis cases) to encoded MALDI-TOF spectra, which increased accuracy and Cohen's kappa to 92.2% and 84.1% respectively. A computational framework integrating protein-protein networks and structural protein information to the machine learning results unveiled the molecular determinants underlying the responsive and unresponsive phenotypes.