Abstract
Effective and timely antibiotic treatment depends on accurate and rapid in silico antimicrobial-resistant (AMR) predictions. Existing statistical rule-based Mycobacterium tuberculosis (MTB) drug resistance prediction methods using bacterial genomic sequencing data often achieve varying results: high accuracy on some antibiotics but relatively low accuracy on others. Traditional machine learning (ML) approaches have been applied to classify drug resistance for MTB and have shown more stable performance. However, there is no study that uses deep learning architecture like Convolutional Neural Network (CNN) on a large and diverse cohort of MTB samples for AMR prediction. We developed 24 binary classifiers of MTB drug resistance status across eight anti-MTB drugs and three different ML algorithms: logistic regression, random forest and 1D CNN using a training dataset of 10,575 MTB isolates collected from 16 countries across six continents, where an extended pan-genome reference was used for detecting genetic features. Our 1D CNN architecture was designed to integrate both sequential and non-sequential features. In terms of F1-scores, 1D CNN models are our best classifiers that are also more accurate and stable than the state-of-the-art rule-based tool Mykrobe predictor (81.1 to 93.8%, 93.7 to 96.2%, 93.1 to 94.8%, 95.9 to 97.2% and 97.1 to 98.2% for ethambutol, rifampicin, pyrazinamide, isoniazid and ofloxacin respectively). We applied filter-based feature selection to find AMR relevant features. All selected variant features are AMR-related ones in CARD database. 78.8% of them are also in the catalogue of MTB mutations that were recently identified as drug resistance-associated ones by WHO. To facilitate ML model development for AMR prediction, we packaged every step into an automated pipeline and shared the source code at https://github.com/KuangXY3/MTB-AMR-classification-CNN.
Highlights
PATRIC Pathosystems Resource Integration Center support vector machine (SVM) Support vector machine whole genome sequencing (WGS) Whole-genome sequencing sequence read archive (SRA) Sequence read archive drug susceptibility test (DST) Drug susceptibility test TP True positive True Negatives (TN) True negative FP False positive FN False negative
To compare the performance of our machine learning (ML) classifiers with a state-of-the-art statistical modeling method Mykrobe predictor, we evaluated the accuracy of Mykrobe predictor on the same dataset[14]
The results showed that our best ML classifiers outperformed the state-of-the-art rule-based method Mykrobe predictor, especially for EMB resistance, and showed more stable accuracy to all the four first-line drugs
Summary
PATRIC Pathosystems Resource Integration Center SVM Support vector machine WGS Whole-genome sequencing SRA Sequence read archive DST Drug susceptibility test TP True positive TN True negative FP False positive FN False negative. There is an urgent need to rapidly identify drug sensitivity profiles of TB, given the fact that culture-based diagnostic tests are usually time-consuming To overcome these restrictions and identify antibiotic resistance more efficiently, researchers use conventional association rule methods to predict antimicrobial r esistance[6]. These methods are based on the identification of variants associated with AMR from whole genome sequencing (WGS) data. The results showed that our best ML classifiers outperformed the state-of-the-art rule-based method Mykrobe predictor, especially for EMB resistance, and showed more stable accuracy to all the four first-line drugs. Our basic 1D CNN architecture didn’t significantly outperform our traditional ML methods LR and RF, there are potential ways to optimize it in the future, e.g., hyperparameter tuning
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
