Abstract
Current machine learning classifiers have successfully been applied to whole-genome sequencing data to identify genetic determinants of antimicrobial resistance (AMR), but they lack causal interpretation. Here we present a metabolic model-based machine learning classifier, named Metabolic Allele Classifier (MAC), that uses flux balance analysis to estimate the biochemical effects of alleles. We apply the MAC to a dataset of 1595 drug-tested Mycobacterium tuberculosis strains and show that MACs predict AMR phenotypes with accuracy on par with mechanism-agnostic machine learning models (isoniazid AUC = 0.93) while enabling a biochemical interpretation of the genotype-phenotype map. Interpretation of MACs for three antibiotics (pyrazinamide, para-aminosalicylic acid, and isoniazid) recapitulates known AMR mechanisms and suggest a biochemical basis for how the identified alleles cause AMR. Extending flux balance analysis to identify accurate sequence classifiers thus contributes mechanistic insights to GWAS, a field thus far dominated by mechanism-agnostic results.
Highlights
Current machine learning classifiers have successfully been applied to whole-genome sequencing data to identify genetic determinants of antimicrobial resistance (AMR), but they lack causal interpretation
The corresponding drug susceptibility status for a strain is described by a binary ‘susceptibility’ or ‘resistance’ phenotype to a particular antibiotic. iEK1011 accounts for 1011 genes (26% of H37Rv) and comprises a metabolic network of 1229 reactions and 998 metabolites
Comparing the gene list between iEK1011 and the genomics dataset, we found that 26% (981/3739) of the total genes and 25% (3310/12,762) of the total variants described by the genetic variant matrix were accounted for by the genome-scale models (GEMs)
Summary
Current machine learning classifiers have successfully been applied to whole-genome sequencing data to identify genetic determinants of antimicrobial resistance (AMR), but they lack causal interpretation. To elucidate AMR mechanisms, researchers have applied machine learning approaches to large-scale genome sequencing and drug-testing datasets for identifying genetic determinants of AMR2–7. While current machine learning approaches have provided a predictive tool for microbial genome-wide association studies (GWAS), such black-box models are incapable of mechanistically interpreting genetic associations. Such a limitation has become increasingly apparent in TB, where numerous experimental studies have shown that AMR-associated genetic variants often reflect network-level metabolic adaptations to antibiotic-induced selection pressures (Supplementary Fig. 1, Supplementary Table 1)[8,9,10,11,12]. This work demonstrates how GEMs can be used directly as an input-output machine learning model to extract both genetic and biochemical network-level insights from microbial GWAS datasets
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
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 callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
