Combining comparative genomic analysis with machine learning reveals some promising diagnostic markers to identify five common pathogenic non-tuberculous mycobacteria

Harmonization of supervised machine learning practices for efficient source attribution of Listeria monocytogenes based on genomic data

BACKGROUND

Genomic data-based machine learning tools are promising for real-time surveillance activities performing source attribution of foodborne bacteria such as Listeria monocytogenes. Given the heterogeneity of machine learning practices, our aim was to identify those influencing the source prediction performance of the usual holdout method combined with the repeated k-fold cross-validation method.

METHODS

A large collection of 1 100 L. monocytogenes genomes with known sources was built according to several genomic metrics to ensure authenticity and completeness of genomic profiles. Based on these genomic profiles (i.e. 7-locus alleles, core alleles, accessory genes, core SNPs and pan kmers), we developed a versatile workflow assessing prediction performance of different combinations of training dataset splitting (i.e. 50, 60, 70, 80 and 90%), data preprocessing (i.e. with or without near-zero variance removal), and learning models (i.e. BLR, ERT, RF, SGB, SVM and XGB). The performance metrics included accuracy, Cohen's kappa, F1-score, area under the curves from receiver operating characteristic curve, precision recall curve or precision recall gain curve, and execution time.

RESULTS

The testing average accuracies from accessory genes and pan kmers were significantly higher than accuracies from core alleles or SNPs. While the accuracies from 70 and 80% of training dataset splitting were not significantly different, those from 80% were significantly higher than the other tested proportions. The near-zero variance removal did not allow to produce results for 7-locus alleles, did not impact significantly the accuracy for core alleles, accessory genes and pan kmers, and decreased significantly accuracy for core SNPs. The SVM and XGB models did not present significant differences in accuracy between each other and reached significantly higher accuracies than BLR, SGB, ERT and RF, in this order of magnitude. However, the SVM model required more computing power than the XGB model, especially for high amount of descriptors such like core SNPs and pan kmers.

CONCLUSIONS

In addition to recommendations about machine learning practices for L. monocytogenes source attribution based on genomic data, the present study also provides a freely available workflow to solve other balanced or unbalanced multiclass phenotypes from binary and categorical genomic profiles of other microorganisms without source code modifications.

Comparative genomic analysis reveals differential genomic characteristics and featured genes between rapid- and slow-growing non-tuberculous mycobacteria

Introduction

Non-tuberculous mycobacteria (NTM) is a major category of environmental bacteria in nature that can be divided into rapidly growing mycobacteria (RGM) and slowly growing mycobacteria (SGM) based on their distinct growth rates. To explore differential molecular mechanisms between RGM and SGM is crucial to understand their survival state, environmental/host adaptation and pathogenicity. Comparative genomic analysis provides a powerful tool for deeply investigating differential molecular mechanisms between them. However, large-scale comparative genomic analysis between RGM and SGM is still uncovered.

Methods

In this study, we screened 335 high-quality, non-redundant NTM genome sequences covering 187 species from 3,478 online NTM genomes, and then performed a comprehensive comparative genomic analysis to identify differential genomic characteristics and featured genes/protein domains between RGM and SGM.

Results

Our findings reveal that RGM has a larger genome size, more genes, lower GC content, and more featured genes/protein domains in metabolism of some main substances (e.g. carbohydrates, amino acids, nucleotides, ions, and coenzymes), energy metabolism, signal transduction, replication, transcription, and translation processes, which are essential for its rapid growth requirements. On the other hand, SGM has a smaller genome size, fewer genes, higher GC content, and more featured genes/protein domains in lipid and secondary metabolite metabolisms and cellular defense mechanisms, which help enhance its genome stability and environmental adaptability. Additionally, orthogroup analysis revealed the important roles of bacterial division and bacteriophage associated genes in RGM and secretion system related genes for better environmental adaptation in SGM. Notably, PCoA analysis of the top 20 genes/protein domains showed precision classification between RGM and SGM, indicating the credibility of our screening/classification strategies.

Discussion

Overall, our findings shed light on differential underlying molecular mechanisms in survival state, adaptation and pathogenicity between RGM and SGM, show the potential for our comparative genomic pipeline to investigate differential genes/protein domains at whole genomic level across different bacterial species on a large scale, and provide an important reference and improved understanding of NTM.

Phylogenomics of nontuberculous mycobacteria respiratory infections in people with cystic fibrosis

Nontuberculous mycobacteria (NTM) can cause severe pulmonary disease in people with cystic fibrosis (pwCF). These infections present unique challenges for diagnosis and treatment, prompting a recent interest in understanding NTM transmission and pathogenesis during chronic infection. Major gaps remain in our knowledge regarding basic pathogenesis, immune evasion strategies, population dynamics, recombination potential, and the evolutionary implications of host and antibiotic pressures of long-term NTM infections in pwCF. Phylogenomic techniques have emerged as an important tool for tracking global patterns of transmission and are beginning to be used to ask fundamental biological questions about adaptation to the host during pathogenesis. In this review, we discuss the burden of NTM lung disease (NTM-LD), highlight the use of phylogenomics in NTM research, and address the clinical implications associated with these studies.

Genomic analysis of Mycobacterium brumae sustains its nonpathogenic and immunogenic phenotype

Mycobacterium brumae is a rapid-growing, non-pathogenic Mycobacterium species, originally isolated from environmental and human samples in Barcelona, Spain. Mycobacterium brumae is not pathogenic and it's in vitro phenotype and immunogenic properties have been well characterized. However, the knowledge of its underlying genetic composition is still incomplete. In this study, we first describe the 4 Mb genome of the M. brumae type strain ATCC 51384T assembling PacBio reads, and second, we assess the low intraspecies variability by comparing the type strain with Illumina reads from three additional strains. Mycobacterium brumae genome is composed of a circular chromosome with a high GC content of 69.2% and containing 3,791 CDSs, 97 pseudogenes, one prophage and no CRISPR loci. Mycobacterium brumae has shown no pathogenic potential in in vivo experiments, and our genomic analysis confirms its phylogenetic position with other non-pathogenic and rapid growing mycobacteria. Accordingly, we determined the absence of virulence-related genes, such as ESX-1 locus and most PE/PPE genes, among others. Although the immunogenic potential of M. brumae was proved to be as high as Mycobacterium bovis BCG, the only mycobacteria licensed to treat cancer, the genomic content of M. tuberculosis T cell and B cell antigens in M. brumae genome is considerably lower than those antigens present in M. bovis BCG genome. Overall, this work provides relevant genomic data on one of the species of the mycobacterial genus with high therapeutic potential.

Comparative genome analysis reveals high-level drug resistance markers in a clinical isolate of Mycobacterium fortuitum subsp. fortuitum MF GZ001

Introduction

Infections caused by non-tuberculosis mycobacteria are significantly worsening across the globe. M. fortuitum complex is a rapidly growing pathogenic species that is of clinical relevance to both humans and animals. This pathogen has the potential to create adverse effects on human healthcare.

Methods

The MF GZ001 clinical strain was collected from the sputum of a 45-year-old male patient with a pulmonary infection. The morphological studies, comparative genomic analysis, and drug resistance profiles along with variants detection were performed in this study. In addition, comparative analysis of virulence genes led us to understand the pathogenicity of this organism.

Results

Bacterial growth kinetics and morphology confirmed that MF GZ001 is a rapidly growing species with a rough morphotype. The MF GZ001 contains 6413573 bp genome size with 66.18 % high G+C content. MF GZ001 possesses a larger genome than other related mycobacteria and included 6156 protein-coding genes. Molecular phylogenetic tree, collinearity, and comparative genomic analysis suggested that MF GZ001 is a novel member of the M. fortuitum complex. We carried out the drug resistance profile analysis and found single nucleotide polymorphism (SNP) mutations in key drug resistance genes such as rpoB, katG, AAC(2')-Ib, gyrA, gyrB, embB, pncA, blaF, thyA, embC, embR, and iniA. In addition, the MF GZ001strain contains mutations in iniA, iniC, pncA, and ribD which conferred resistance to isoniazid, ethambutol, pyrazinamide, and para-aminosalicylic acid respectively, which are not frequently observed in rapidly growing mycobacteria. A wide variety of predicted putative potential virulence genes were found in MF GZ001, most of which are shared with well-recognized mycobacterial species with high pathogenic profiles such as M. tuberculosis and M. abscessus.

Discussion

Our identified novel features of a pathogenic member of the M. fortuitum complex will provide the foundation for further investigation of mycobacterial pathogenicity and effective treatment.

Mycobacterium abscessus: insights from a bioinformatic perspective

Mycobacterium abscessus is a nontuberculous mycobacterium, associated with broncho-pulmonary infections in individuals suffering from cystic fibrosis, bronchiectasis, and pulmonary diseases. The risk factors for transmission include biofilms, contaminated water resources, fomites, and infected individuals. M. abscessus is extensively resistant to antibiotics. To date, there is no vaccine and combination antibiotic therapy is followed. However, drug toxicities, low cure rates, and high cost of treatment make it imperfect. Over the last 20 years, bioinformatic studies on M. abscessus have advanced our understanding of the pathogen. This review integrates knowledge from the analysis of genomes, microbiomes, genomic variations, phylogeny, proteome, transcriptome, secretome, antibiotic resistance, and vaccine design to further our understanding. The utility of genome-based studies in comprehending disease progression, surveillance, tracing transmission routes, and epidemiological outbreaks on a global scale has been highlighted. Furthermore, this review underlined the importance of using computational methodologies for pinpointing factors responsible for pathogen survival and resistance. We reiterate the significance of interdisciplinary research to fight M. abscessus. In a nutshell, the outcome of computational studies can go a long way in creating novel therapeutic avenues to control M. abscessus mediated pulmonary infections.