Using random forest to predict antimicrobial minimum inhibitory concentrations of nontyphoidal Salmonella in Taiwan

Abstract

Antimicrobial resistance (AMR) is a global health issue and surveillance of AMR can be useful for understanding AMR trends and planning intervention strategies. Salmonella, widely distributed in food-producing animals, has been considered the first priority for inclusion in the AMR surveillance program by the World Health Organization (WHO). Recent advances in rapid and affordable whole-genome sequencing (WGS) techniques lead to the emergence of WGS as a one-stop test to predict the antimicrobial susceptibility. Since the variation of sequencing and minimum inhibitory concentration (MIC) measurement methods could result in different results, this study aimed to develop WGS-based random forest models for predicting MIC values of 24 drugs using data generated from the same laboratories in Taiwan. The WGS data have been transformed as a feature vector of 10-mers for machine learning. Based on rigorous validation and independent tests, a good performance was obtained with an average mean absolute error (MAE) less than 1 for both validation and independent test. Feature selection was then applied to identify top-ranked 10-mers that can further improve the prediction performance. For surveillance purposes, the genome sequence-based machine learning methods could be utilized to monitor the difference between predicted and experimental MIC, where a large difference might be worthy of investigation on the emerging genomic determinants.

Predicting Salmonella MIC and Deciphering Genomic Determinants of Antibiotic Resistance and Susceptibility

“…Considering potential sequencing variations, it has been recommended to construct prediction models based on well-controlled experiments using WGS data sourced from the same laboratories [ 34 ]. The study achieved an average prediction accuracy of 92% for 24 antibiotics, with 321 WGS as predictors.…”

Machine learning and feature extraction for rapid antimicrobial resistance prediction of Acinetobacter baumannii from whole-genome sequencing data

“…The developed models exhibited predictive capabilities. However, the presence of high-dimensional feature vectors may potentially impact machine learning performance and increase execution time ( Wang et al, 2023 ). We further analysed the best models of the 13 antimicrobial agents and used the feature and importance function estimator integrated within RF and SVM-linear to summarise and rank the importance of the 11-mers features.…”

“…Termed essential agreement (EA) and category agreement (CA), receiver operating characteristic (ROC) curves, and area under the curve (AUC) values were also used to judge the predictive performance of the models. For the final classification results based on clinical breakpoints, the recall/sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), major error (ME), and very major error (VME) were also calculated to evaluate model performance as follows: Recall/Sensitivity = TP/(TP + FN); Specificity = TN/(TN + FP); PPV = TP/(TP + FP); NPV = TN/(TN + FN); ME = FP/(TN + FP); VME = FN/(TP + FN), where TP, FN, TN, and FP represent true positives, false negatives, true negatives, and false positives, respectively ( Wang et al, 2023 ).…”

“…Pre-existing knowledge is indispensable in obtaining reliable results ( Maguire et al, 2019 ). A limited number of studies have focused on applying WGS and AMR phenotype data without prior information to train machine-learning models for AMR prediction of non-typhoidal Salmonella , Escherichia coli , and Klebsiella pneumoniae ( Nguyen et al, 2018 ; Maguire et al, 2019 ; Nguyen et al, 2019 ; Humphries et al, 2023 ; Wang et al, 2023 ). Although this reference-free approach necessitates a substantial volume of genomes, it is unbiased and thus facilitates the detection of novel genomic features implicated in AMR ( Drouin et al, 2016 ).…”

Quantitative prediction of disinfectant tolerance inListeria monocytogenesusing whole genome sequencing and machine learning

“…Our analysis showed promising results for direct MIC prediction, and we identified RF and GBT models as top performers with cross-validated MSE scores on log 2 transformed MIC values of 0.25±0.03 and 0.28±0.05, respectively. Even though only a limited number of studies have explored MIC prediction of disinfectants from genomic data, making comparisons difficult, tree-based models show good performance and are commonly used for similar prediction tasks [15,16,38,39].…”