Dataset Information

Developing an in silico minimum inhibitory concentration panel test for Klebsiella pneumoniae.

ABSTRACT: Antimicrobial resistant infections are a serious public health threat worldwide. Whole genome sequencing approaches to rapidly identify pathogens and predict antibiotic resistance phenotypes are becoming more feasible and may offer a way to reduce clinical test turnaround times compared to conventional culture-based methods, and in turn, improve patient outcomes. In this study, we use whole genome sequence data from 1668 clinical isolates of Klebsiella pneumoniae to develop a XGBoost-based machine learning model that accurately predicts minimum inhibitory concentrations (MICs) for 20 antibiotics. The overall accuracy of the model, within ±1 two-fold dilution factor, is 92%. Individual accuracies are ?90% for 15/20 antibiotics. We show that the MICs predicted by the model correlate with known antimicrobial resistance genes. Importantly, the genome-wide approach described in this study offers a way to predict MICs for isolates without knowledge of the underlying gene content. This study shows that machine learning can be used to build a complete in silico MIC prediction panel for K. pneumoniae and provides a framework for building MIC prediction models for other pathogenic bacteria.

SUBMITTER: Nguyen M

PROVIDER: S-EPMC5765115 | biostudies-literature | 2018 Jan

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Developing an in silico minimum inhibitory concentration panel test for Klebsiella pneumoniae.

Nguyen Marcus M Brettin Thomas T Long S Wesley SW Musser James M JM Olsen Randall J RJ Olson Robert R Shukla Maulik M Stevens Rick L RL Xia Fangfang F Yoo Hyunseung H Davis James J JJ

Scientific reports 20180111 1

Antimicrobial resistant infections are a serious public health threat worldwide. Whole genome sequencing approaches to rapidly identify pathogens and predict antibiotic resistance phenotypes are becoming more feasible and may offer a way to reduce clinical test turnaround times compared to conventional culture-based methods, and in turn, improve patient outcomes. In this study, we use whole genome sequence data from 1668 clinical isolates of Klebsiella pneumoniae to develop a XGBoost-based machi ...[more]

PMID: 29323230

Similar Datasets

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

Project description:To illustrate the genomic and drug resistance traits of the Klebsiella pneumoniae Kpn_XM9, which harbors a transposon (Tn) As1 and was barely susceptible to ceftazidime-avibactam (CZA). Whole-genome sequencing, gene deletion, antimicrobial susceptibility, and conjugation tests were carried out to illustrate the traits of Kpn_XM9. As confirmed by whole-genome sequencing, the Kpn_XM9 harbored a 5,523,536 bp chromosome and five plasmids with lengths being 128,129, 196,512, 84,812, 43,695, and 5,596 bp, respectively. Plasmid p1_Kpn_XM9 (128,219 bp) contained four resistance genes, blaCTX-M-65, blaTEM-1B, rmtB, and two copies of blaKPC-2. Genes blaKPC-2 were bracketed by ISKpn17 and ISKpn16 within a new composite Tn3-like TnAs1. The two tandem repeats, positioned opposite each other, were spaced 93,447 bp apart in p1_Kpn_XM9. Kpn_XM9 belonged to K64 and sequence type (ST) 11. The Kpn_XM9 was resistant to amikacin, aztreonam, ticarcillin/clavulanic acid, piperacillin/tazobactam, ceftazidime, cefepime, imipenem, meropenem, tobramycin, ciprofloxacin, levofloxacin, doxycycline, minocycline, tigecycline, colistin, and trimethoprim/sulfamethoxazole; it was barely susceptible to CZA with a minimum inhibitory concentration of 8/4 µg/mL, which declined to 2/4 µg/mL after a 18,555 bp nucleotide was knocked out and one copy of blaKPC-2 was sustained on p1_Kpn_XM9. Kpn_XM9 had virulence genes encoding Types 1 and 3 fimbriae, four siderophores, and capsular polysaccharide anchoring protein but no genes upregulating capsular polysaccharide synthesis. The Kpn_XM9 presented a classical phenotype with extreme drug resistance. The emergence of double copies of blaKPC-2 in a single plasmid from the predominant ST11 K. pneumoniae represents a new therapeutic challenge.IMPORTANCEWith the wide use of ceftazidime-avibactam against carbapenem-resistant organisms, its resistance is increasingly documented; among the corresponding resistance mechanisms, mutations of blaKPC-2 or blaKPC-3 into other subtypes are dominant to date. However, more copies of blaKPC-2 may also greatly increase the minimum inhibitory concentration of ceftazidime-avibactam, which could be conferred by transposon As1 and insertion sequence 26 and should be of concern.

Project description:Minimal inhibitory concentration (MIC) is defined as the lowest concentration of an antimicrobial agent that can inhibit the visible growth of a particular microorganism after overnight incubation. Clinically, antibiotic doses for specific infections are determined according to the fraction of MIC. Therefore, credible assessment of MICs will provide a physician valuable information on the choice of therapeutic strategy. Early and precise usage of antibiotics is the key to an infection therapy. Compared with the traditional culture-based method, the approach of whole genome sequencing to identify MICs can shorten the experimental time, thereby improving clinical efficacy. Klebsiella pneumoniae is one of the most significant members of the genus Klebsiella in the Enterobacteriaceae family and also a common non-social pathogen. Meropenem is a broad-spectrum antibacterial agent of the carbapenem family, which can produce antibacterial effects of most Gram-positive and -negative bacteria. In this study, we used single-nucleotide polymorphism (SNP) information and nucleotide k-mers count based on metagenomic data to predict MICs of meropenem against K. pneumoniae. Then, features of 110 sequenced K. pneumoniae genome data were combined and modeled with XGBoost algorithm and deep neural network (DNN) algorithm to predict MICs. We first use the XGBoost classification model and the XGBoost regression model. After five runs, the average accuracy of the test set was calculated. The accuracy of using nucleotide k-mers to predict MICs of the XGBoost classification model and XGBoost regression model was 84.5 and 89.1%. The accuracy of SNP in predicting MIC was 80 and 81.8%, respectively. The results show that XGBoost regression is better than XGBoost classification in both nucleotide k-mers and SNPs to predict MICs. We further selected 40 nucleotide k-mers and 40 SNPs with the highest correlation with MIC values as features to retrain the XGBoost regression model and DNN regression model. After 100 and 1,000 runs, the results show that the accuracy of the two models was improved. The accuracy of the XGBoost regression model for k-mers, SNPs, and k-mers & SNPs was 91.1, 85.2, and 91.3%, respectively. The accuracy of the DNN regression model was 91.9, 87.1, and 91.8%, respectively. Through external verification, some of the selected features were found to be related to drug resistance.

Project description:Gram-negative bacteria possess stress responses to maintain the integrity of the cell envelope. Stress sensors monitor outer membrane permeability, envelope protein folding, and energization of the inner membrane. The systems used by Gram-negative bacteria to sense and combat stress resulting from disruption of the peptidoglycan layer are not well characterized. The peptidoglycan layer is a single molecule that completely surrounds the cell and ensures its structural integrity. During cell growth new peptidoglcyan subunits are incorporated into the peptidoglycan layer by a series of enzymes called the penicillin-binding proteins (PBPs). To explore how Gram-negative bacteria respond to peptidoglycan stress, global gene expression analysis was used to identify Escherichia coli stress responses activated following inhibition of specific PBPs by the β-lactam antibiotics mecillinam and cefsulodin. Inhibition of PBPs with different roles in peptidoglycan synthesis has different consequences for cell morphology and viability, suggesting that not all perturbations to the peptidoglycan layer generate equivalent stresses. We demonstrate that inhibition of different PBPs resulted in both shared and unique stress responses. The regulation of capsular synthesis (Rcs) phosphorelay was activated by inhibition of all of the PBPs tested. Furthermore, we show that activation of the Rcs phosphorelay increased survival in the presence of these antibiotics, independently of capsule synthesis. Both activation of the phosphorelay and survival required signal transduction via the outer membrane lipoprotein RcsF and the response regulator RcsB. We propose that the Rcs pathway responds to peptidoglycan damage and contributes to the intrinsic resistance of E. coli to β-lactam antibiotics. We used microarrays to identify changes in gene expression resulting from treatment of Escherichia coli with the β-lactam antibiotics cefsulodin, mecillinam, or the combination. Keywords: Dose response, stress response

Project description:Background:Vancomycin, the first line antibiotic for methicillin-resistant Staphylococcus aureus (MRSA) bacteremia, is often administered inappropriately when MIC is greater than 2 µg/mL, including 'susceptible' strains. This study assessed the discordance of vancomycin minimum inhibitory concentration (MIC) for methicillin-resistant Staphylococcus aureus (MRSA). Methods:In total, 229 MRSA isolates from blood cultures collected between 2009 and 2015 at a tertiary hospital in Taiwan were examined. The MICs of vancomycin were measured using Vitek 2, E-test, and standard broth microdilution at the level of 2 µg/mL. Results:The geometric mean of the MICs of hospital-acquired MRSA was higher than that of community-acquired MRSA (P < 0.001), with the exact agreement rates (with broth microdilution) at 2 µg/mL being 53.6% in Vitek 2 and 86.7% in E-test. Overall, E-test (98.1%) had more categorical accordance than did Vitek 2 (94.0%; P = 0.026). Vitek 2 had a tendency to overestimate MRSA in high-MIC isolates, whereas E-test inclined underestimation in low-MIC isolates. Surprisingly, the discordance rates of MRSA vancomycin MICs were higher in hospital-acquired isolates (13.3%-17.0%) than in community-acquired isolates (6.2%-7.0%). Conclusion:The Infectious Diseases Society of America recommends the use of alternative antimicrobial agents when vancomycin MIC is ? 2 µg/mL; in this study, only 53.6% of the isolates tested using Vitek 2 showed a high MIC in the broth microdilution method. Accurate identification of the resistance profile is a key component of antimicrobial stewardship programs. Therefore, to reduce inappropriate antibiotic use and mitigate the emergence of resistant strains, we recommend using complementary tests such as E-test or Broth microdilution to verify the MIC before administering second-line antibiotics. Strengths:(1) We compared the categorical agreement between different methods measuring MRSA MICs level. (2) Physicians should incorporate this information and consider a complementary test to verify the appropriateness of the decision of shifting vancomycin to second-line antibiotic treatment to improve patients' prognosis. (3) MRSA-vancomycin MICs at a cutoff of 2 µg/mL obtained using Vitek II exhibited a higher sensitivity level and negative predictive value than those obtained using E-test in the prediction of categorical agreement with standard broth microdilution. Limitation:(1) Our research was based on a single hospital-based study. (2) The MRSA strains in this study were stored for more than 12 months after isolation. (3) We did not collect information on clinical prognosis.

Dataset Information

Developing an in silico minimum inhibitory concentration panel test for Klebsiella pneumoniae.

Publications

Developing an in silico minimum inhibitory concentration panel test for Klebsiella pneumoniae.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets