Optimising machine learning prediction of minimum inhibitory concentrations in Klebsiella pneumoniae DOI Creative Commons
Gherard Batisti Biffignandi, Leonid Chindelevitch, Marta Corbella

et al.

Microbial Genomics, Journal Year: 2024, Volume and Issue: 10(3)

Published: March 26, 2024

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

Language: Английский

Machine Learning for Antimicrobial Resistance Prediction: Current Practice, Limitations, and Clinical Perspective DOI
Jee In Kim, Finlay Maguire, Kara K. Tsang

et al.

Clinical Microbiology Reviews, Journal Year: 2022, Volume and Issue: 35(3)

Published: May 25, 2022

Antimicrobial resistance (AMR) is a global health crisis that poses great threat to modern medicine. Effective prevention strategies are urgently required slow the emergence and further dissemination of AMR. Given availability data sets encompassing hundreds or thousands pathogen genomes, machine learning (ML) increasingly being used predict different antibiotics in pathogens based on gene content genome composition. A key objective this work advocate for incorporation ML into front-line settings but also highlight refinements necessary safely confidently incorporate these methods. The question what not trivial given existence quantitative qualitative laboratory measures models typically treat genes as independent predictors, with no consideration structural functional linkages; they may be accurate when new mutational variants known AMR emerge. Finally, have technology trusted by end users public settings, need transparent explainable ensure basis prediction clear. We strongly next set AMR-ML studies should focus refinement limitations able bridge gap diagnostic implementation.

Language: Английский

Citations

94
The phylogenetic landscape and nosocomial spread of the multidrug-resistant opportunist Stenotrophomonas maltophilia DOI Creative Commons
Matthias I. Gröschel, Conor J. Meehan, Ivan Barilar

et al.

Nature Communications, Journal Year: 2020, Volume and Issue: 11(1)

Published: April 27, 2020

Abstract Recent studies portend a rising global spread and adaptation of human- or healthcare-associated pathogens. Here, we analyse an international collection the emerging, multidrug-resistant, opportunistic pathogen Stenotrophomonas maltophilia from 22 countries to infer population structure clonality at level. We show that S. complex is divided into 23 monophyletic lineages, most which harbour strains all degrees human virulence. Lineage Sm6 comprises highest rate human-associated strains, linked key virulence resistance genes. Transmission analysis identifies potential outbreak events genetically closely related isolated within days weeks in same hospitals.

Language: Английский

Citations

104
Improved Prediction of Bacterial Genotype-Phenotype Associations Using Interpretable Pangenome-Spanning Regressions DOI Creative Commons
John A. Lees, The Tien Mai, Marco Galardini

et al.

mBio, Journal Year: 2020, Volume and Issue: 11(4)

Published: July 6, 2020

Being able to identify the genetic variants responsible for specific bacterial phenotypes has been goal of genetics since its inception and is fundamental our current level understanding bacteria. This identification based primarily on painstaking experimentation, but availability large data sets whole genomes with associated phenotype metadata promises revolutionize this approach, not least important clinical that are amenable laboratory analysis. These models phenotype-genotype association can in future be used rapid prediction clinically such as antibiotic resistance virulence by rapid-turnaround or point-of-care tests. However, despite much effort being put into adapting genome-wide study (GWAS) approaches cope bacterium-specific problems, strong population structure horizontal gene exchange, yet optimal. We describe a method advances methodology both generation portable models.

Language: Английский

Citations

101
Current Uses and Future Perspectives of Genomic Technologies in Clinical Microbiology DOI Creative Commons
Irene Bianconi,

Richard Aschbacher,

Elisabetta Pagani

et al.

Antibiotics, Journal Year: 2023, Volume and Issue: 12(11), P. 1580 - 1580

Published: Oct. 30, 2023

Recent advancements in sequencing technology and data analytics have led to a transformative era pathogen detection typing. These developments not only expedite the process, but also render it more cost-effective. Genomic analyses of infectious diseases are swiftly becoming standard for analysis control. Additionally, national surveillance systems can derive substantial benefits from genomic data, as they offer profound insights into epidemiology emergence antimicrobial-resistant strains. Antimicrobial resistance (AMR) is pressing global public health issue. While clinical laboratories traditionally relied on culture-based antimicrobial susceptibility testing, integration AMR holds immense promise. Genomic-based furnish swift, consistent, highly accurate predictions phenotypes specific strains or populations, all while contributing invaluable surveillance. Moreover, genome assumes pivotal role investigation hospital outbreaks. It aids identification infection sources, unveils genetic connections among isolates, informs strategies The One Health initiative, with its focus intricate interconnectedness humans, animals, environment, seeks develop comprehensive approaches disease surveillance, control, prevention. When integrated epidemiological systems, forecast expansion bacterial populations species transmissions. Consequently, this provides evolution relationships pathogens, hosts, environment.

Language: Английский

Citations

28
Forest and Trees: Exploring Bacterial Virulence with Genome-wide Association Studies and Machine Learning DOI Creative Commons
Jonathan P. Allen, Evan S. Snitkin, Nathan B. Pincus

et al.

Trends in Microbiology, Journal Year: 2021, Volume and Issue: 29(7), P. 621 - 633

Published: Jan. 14, 2021

Language: Английский

Citations

54
Increased power from conditional bacterial genome-wide association identifies macrolide resistance mutations in Neisseria gonorrhoeae DOI Creative Commons
C. Kevin, Tatum D. Mortimer,

Marissa A. Duckett

et al.

Nature Communications, Journal Year: 2020, Volume and Issue: 11(1)

Published: Oct. 23, 2020

Abstract The emergence of resistance to azithromycin complicates treatment Neisseria gonorrhoeae , the etiologic agent gonorrhea. Substantial remains unexplained after accounting for known mutations. Bacterial genome-wide association studies (GWAS) can identify novel genes but must control genetic confounders while maintaining power. Here, we show that compared single-locus GWAS, conducting GWAS conditioned on mutations reduces number false positives and identifies a G70D mutation in RplD 50S ribosomal protein L4 as significantly associated with increased ( p -value = 1.08 × 10 −11 ). We experimentally confirm our results demonstrate other macrolide binding site are prevalent (present 5.42% 4850 isolates) widespread (identified 21/65 countries across two decades). Overall, findings utility conditional associations improving performance microbial advance understanding basis resistance.

Language: Английский

Citations

53
Molecular detection and characterization of foodborne bacteria: Recent progresses and remaining challenges DOI Creative Commons
Joshua Hadi, Delphine Rapp, Sharduli Dhawan

et al.

Comprehensive Reviews in Food Science and Food Safety, Journal Year: 2023, Volume and Issue: 22(3), P. 2433 - 2464

Published: April 11, 2023

The global food demand is expected to increase in the coming years, along with challenges around climate change and security. Concomitantly, safety risks, particularly those related bacterial pathogens, may also increase. Thus, sector needs innovate rise this challenge. Here, we discuss recent advancements molecular techniques that can be deployed within various foodborne bacteria surveillance systems across settings. To start with, provide updates on nucleic acid-based detection, a focus polymerase chain reaction (PCR)-based technologies loop-mediated isothermal amplification (LAMP). These include descriptions of novel genetic markers for several progresses multiplex PCR droplet digital PCR. next section provides an overview development clustered regularly interspaced short palindromic repeats (CRISPR) CRISPR-associated (Cas) proteins systems, such as CRISPR-Cas9, CRISPR-Cas12a, CRISPR-Cas13a, tools enhanced sensitive specific detection pathogens. final describes utilizations whole genome sequencing accurate characterization bacteria, ranging from epidemiological model-based predictions phenotypic traits through genome-wide association studies or machine learning.

Language: Английский

Citations

18
From genotype to phenotype: computational approaches for inferring microbial traits relevant to the food industry DOI Creative Commons
Signe Tang Karlsen, Martin Holm Rau, Benjamín J. Sánchez

et al.

FEMS Microbiology Reviews, Journal Year: 2023, Volume and Issue: 47(4)

Published: June 7, 2023

Abstract When selecting microbial strains for the production of fermented foods, various phenotypes need to be taken into account achieve target product characteristics, such as biosafety, flavor, texture, and health-promoting effects. Through continuous advances in sequencing technologies, whole-genome sequences increasing quality can now obtained both cheaper faster, which increases relevance genome-based characterization phenotypes. Prediction from genome makes it possible quickly screen large strain collections silico identify candidates with desirable traits. Several relevant foods predicted using knowledge-based approaches, leveraging our existing understanding genetic molecular mechanisms underlying those In absence this knowledge, data-driven approaches applied estimate genotype–phenotype relationships based on experimental datasets. Here, we review computational methods that implement knowledge- phenotype prediction, well combine elements approaches. Furthermore, provide examples how these have been industrial biotechnology, special focus food industry.

Language: Английский

Citations

18
Pathogen-associated gene discovery workflows for novel antivirulence therapeutic development DOI Creative Commons
Wing Yin Venus Lau,

Patrick K. Taylor,

Fiona S. L. Brinkman

et al.

EBioMedicine, Journal Year: 2023, Volume and Issue: 88, P. 104429 - 104429

Published: Jan. 9, 2023

Novel therapeutics to manage bacterial infections are urgently needed as the impact and prevalence of antimicrobial resistance (AMR) grows. Antivirulence an alternative approach antibiotics that aim attenuate virulence rather than target essential functions, while minimizing microbiota perturbation risk AMR development. Beyond known factors, pathogen-associated genes (PAGs; found only in pathogens date) may play important role or host association. Many identified PAGs encode uncharacterized hypothetical proteins represent untapped wealth novel drug targets. Here, we review current advances antivirulence research development, including PAG identification, provide a comprehensive workflow from discovery targets discovery. We highlight importance integrating bioinformatic/genomic-based methods for factor discovery, coupled with experimental characterization, into existing screening platforms develop effective drugs.

Language: Английский

Citations

17
Genome wide association study of Escherichia coli bloodstream infection isolates identifies genetic determinants for the portal of entry but not fatal outcome DOI Creative Commons
Érick Denamur, Bénédicte Condamine,

Marina Esposito‐Farèse

et al.

PLoS Genetics, Journal Year: 2022, Volume and Issue: 18(3), P. e1010112 - e1010112

Published: March 24, 2022

Escherichia coli is an important cause of bloodstream infections (BSI), which concern given its high mortality and increasing worldwide prevalence. Finding bacterial genetic variants that might contribute to patient death interest better understand infection progression implement diagnostic methods specifically look for those factors. E. samples isolated from patients with BSI are ideal dataset systematically search variants, as long the influence host factors such comorbidities taken into account. Here we performed a genome-wide association study (GWAS) using data 912 hospitals in Paris, France. We looked associations between three outcomes (death at 28 days, septic shock admission intensive care unit), well two portals entry (urinary digestive tract), various clinical variables each account did not find any outcomes, potentially confirming strong influencing course BSI; however found papGII operon entrance through urinary tract, demonstrates power GWAS when applied actual data. Despite lack estimate sample size by one order magnitude could lead discovery some putative causal variants. Given wide adoption genome sequencing isolates, sizes may be soon available.

Language: Английский

Citations

25