Feature selection and aggregation for antibiotic resistance GWAS in Mycobacterium tuberculosis: a comparative study

IntroductionDrug resistance (DR) of pathogens remains a global healthcare concern. In contrast to other bacteria, acquiring mutations in the core genome is the main mechanism of drug resistance for Mycobacterium tuberculosis (MTB). For some antibiotics, the resistance of a particular isolate can be...

Full description

Saved in:
Bibliographic Details
Main Authors: Kirill Reshetnikov, Daria Bykova, Konstantin Kuleshov, Konstantin Chukreev, Egor Guguchkin, Alexey Neverov, Gennady Fedonin
Format: Article
Language:English
Published: Frontiers Media S.A. 2025-06-01
Series:Frontiers in Microbiology
Subjects:
Online Access:https://www.frontiersin.org/articles/10.3389/fmicb.2025.1586476/full
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850117850621018112
author Kirill Reshetnikov
Daria Bykova
Daria Bykova
Konstantin Kuleshov
Konstantin Kuleshov
Konstantin Chukreev
Egor Guguchkin
Alexey Neverov
Gennady Fedonin
author_facet Kirill Reshetnikov
Daria Bykova
Daria Bykova
Konstantin Kuleshov
Konstantin Kuleshov
Konstantin Chukreev
Egor Guguchkin
Alexey Neverov
Gennady Fedonin
author_sort Kirill Reshetnikov
collection DOAJ
description IntroductionDrug resistance (DR) of pathogens remains a global healthcare concern. In contrast to other bacteria, acquiring mutations in the core genome is the main mechanism of drug resistance for Mycobacterium tuberculosis (MTB). For some antibiotics, the resistance of a particular isolate can be reliably predicted by identifying specific mutations, while for other antibiotics the knowledge of resistance mechanisms is limited. Statistical machine learning (ML) methods are used to infer new genes implicated in drug resistance leveraging large collections of isolates with known whole-genome sequences and phenotypic states for different drugs. However, high correlations between the phenotypic states for commonly used drugs complicate the inference of true associations of mutations with drug phenotypes by ML approaches.MethodsRecently, several new methods have been developed to select a small subset of reliable predictors of the dependent variable, which may help reduce the number of spurious associations identified. In this study, we evaluated several such methods, namely, logistic regression with different regularization penalty functions, a recently introduced algorithm for solving the best-subset selection problem (ABESS) and “Hungry, Hungry SNPos” (HHS) a heuristic algorithm specifically developed to identify resistance-associated genetic variants in the presence of resistance co-occurrence. We assessed their ability to select known causal mutations for resistance to a specific drug while avoiding the selection of mutations in genes associated with resistance to other drugs, thus we compared selected ML models for their applicability for MTB genome wide association studies.Results and discussionIn our analysis, ABESS significantly outperformed the other methods, selecting more relevant sets of mutations. Additionally, we demonstrated that aggregating rare mutations within protein-coding genes into markers indicative of changes in PFAM domains improved prediction quality, and these markers were predominantly selected by ABESS, suggesting their high informativeness. However, ABESS yielded lower prediction accuracy compared to logistic regression methods with regularization.
format Article
id doaj-art-acbb5aae577941fb954cc399eb3aece1
institution OA Journals
issn 1664-302X
language English
publishDate 2025-06-01
publisher Frontiers Media S.A.
record_format Article
series Frontiers in Microbiology
spelling doaj-art-acbb5aae577941fb954cc399eb3aece12025-08-20T02:36:00ZengFrontiers Media S.A.Frontiers in Microbiology1664-302X2025-06-011610.3389/fmicb.2025.15864761586476Feature selection and aggregation for antibiotic resistance GWAS in Mycobacterium tuberculosis: a comparative studyKirill Reshetnikov0Daria Bykova1Daria Bykova2Konstantin Kuleshov3Konstantin Kuleshov4Konstantin Chukreev5Egor Guguchkin6Alexey Neverov7Gennady Fedonin8Central Research Institute of Epidemiology, Moscow, RussiaCentral Research Institute of Epidemiology, Moscow, RussiaFaculty of Bioengineering and Bioinformatics, Moscow State University, Moscow, RussiaCentral Research Institute of Epidemiology, Moscow, RussiaFederal State Budget Scientific Institution “Federal Scientific Center VIEV”, Moscow, RussiaCentral Research Institute of Epidemiology, Moscow, RussiaUfa Federal Research Center of the Russian Academy of Sciences, Ufa, RussiaFaculty of Computer Science/Institute of Artificial Intelligence and Digital Sciences/International Laboratory for Statistical and Computational Genomics, HSE University, Moscow, RussiaCentral Research Institute of Epidemiology, Moscow, RussiaIntroductionDrug resistance (DR) of pathogens remains a global healthcare concern. In contrast to other bacteria, acquiring mutations in the core genome is the main mechanism of drug resistance for Mycobacterium tuberculosis (MTB). For some antibiotics, the resistance of a particular isolate can be reliably predicted by identifying specific mutations, while for other antibiotics the knowledge of resistance mechanisms is limited. Statistical machine learning (ML) methods are used to infer new genes implicated in drug resistance leveraging large collections of isolates with known whole-genome sequences and phenotypic states for different drugs. However, high correlations between the phenotypic states for commonly used drugs complicate the inference of true associations of mutations with drug phenotypes by ML approaches.MethodsRecently, several new methods have been developed to select a small subset of reliable predictors of the dependent variable, which may help reduce the number of spurious associations identified. In this study, we evaluated several such methods, namely, logistic regression with different regularization penalty functions, a recently introduced algorithm for solving the best-subset selection problem (ABESS) and “Hungry, Hungry SNPos” (HHS) a heuristic algorithm specifically developed to identify resistance-associated genetic variants in the presence of resistance co-occurrence. We assessed their ability to select known causal mutations for resistance to a specific drug while avoiding the selection of mutations in genes associated with resistance to other drugs, thus we compared selected ML models for their applicability for MTB genome wide association studies.Results and discussionIn our analysis, ABESS significantly outperformed the other methods, selecting more relevant sets of mutations. Additionally, we demonstrated that aggregating rare mutations within protein-coding genes into markers indicative of changes in PFAM domains improved prediction quality, and these markers were predominantly selected by ABESS, suggesting their high informativeness. However, ABESS yielded lower prediction accuracy compared to logistic regression methods with regularization.https://www.frontiersin.org/articles/10.3389/fmicb.2025.1586476/fullMycobacterium tuberculosisantimicrobial drug resistancefeature selectionmachine learningPFAM domains
spellingShingle Kirill Reshetnikov
Daria Bykova
Daria Bykova
Konstantin Kuleshov
Konstantin Kuleshov
Konstantin Chukreev
Egor Guguchkin
Alexey Neverov
Gennady Fedonin
Feature selection and aggregation for antibiotic resistance GWAS in Mycobacterium tuberculosis: a comparative study
Frontiers in Microbiology
Mycobacterium tuberculosis
antimicrobial drug resistance
feature selection
machine learning
PFAM domains
title Feature selection and aggregation for antibiotic resistance GWAS in Mycobacterium tuberculosis: a comparative study
title_full Feature selection and aggregation for antibiotic resistance GWAS in Mycobacterium tuberculosis: a comparative study
title_fullStr Feature selection and aggregation for antibiotic resistance GWAS in Mycobacterium tuberculosis: a comparative study
title_full_unstemmed Feature selection and aggregation for antibiotic resistance GWAS in Mycobacterium tuberculosis: a comparative study
title_short Feature selection and aggregation for antibiotic resistance GWAS in Mycobacterium tuberculosis: a comparative study
title_sort feature selection and aggregation for antibiotic resistance gwas in mycobacterium tuberculosis a comparative study
topic Mycobacterium tuberculosis
antimicrobial drug resistance
feature selection
machine learning
PFAM domains
url https://www.frontiersin.org/articles/10.3389/fmicb.2025.1586476/full
work_keys_str_mv AT kirillreshetnikov featureselectionandaggregationforantibioticresistancegwasinmycobacteriumtuberculosisacomparativestudy
AT dariabykova featureselectionandaggregationforantibioticresistancegwasinmycobacteriumtuberculosisacomparativestudy
AT dariabykova featureselectionandaggregationforantibioticresistancegwasinmycobacteriumtuberculosisacomparativestudy
AT konstantinkuleshov featureselectionandaggregationforantibioticresistancegwasinmycobacteriumtuberculosisacomparativestudy
AT konstantinkuleshov featureselectionandaggregationforantibioticresistancegwasinmycobacteriumtuberculosisacomparativestudy
AT konstantinchukreev featureselectionandaggregationforantibioticresistancegwasinmycobacteriumtuberculosisacomparativestudy
AT egorguguchkin featureselectionandaggregationforantibioticresistancegwasinmycobacteriumtuberculosisacomparativestudy
AT alexeyneverov featureselectionandaggregationforantibioticresistancegwasinmycobacteriumtuberculosisacomparativestudy
AT gennadyfedonin featureselectionandaggregationforantibioticresistancegwasinmycobacteriumtuberculosisacomparativestudy