An optimized genomic VCF workflow for precise identification of Mycobacterium tuberculosis cluster from cross-platform whole genome sequencing data

Risk for Prison-to-Community Tuberculosis Transmission, Thailand, 2017-2020

To determine contributions of previously incarcerated persons to tuberculosis (TB) transmission in the community, we performed a healthcare facility-based cohort study of TB patients in Thailand during 2017-2020. We used whole-genome sequencing of Mycobacterium tuberculosis isolates from patients to identify genotypic clusters and assess the association between previous incarceration and TB transmission in the community. We identified 4 large genotype clusters (>10 TB patients/cluster); 28% (14/50) of the patients in those clusters were formerly incarcerated. Formerly incarcerated TB patients were more likely than nonincarcerated patients to be included in large clusters. TB patients within the large genotype clusters were geographically dispersed throughout Chiang Rai Province. Community TB transmission in the community was associated with the presence of formerly incarcerated individuals in Thailand. To reduce the risk for prison-to-community transmission, we recommend TB screening at the time of entry and exit from prisons and follow-up screening in the community.

EfGD: the Erianthus fulvus genome database

Erianthus fulvus (TaxID: 154759) is a valuable germplasm resource in sugarcane breeding and research and has excellent agronomic traits, such as drought resistance, cold resistance, barren tolerance and high brix. With a stable chromosome number (2n = 20) and a small genome (0.9 Gb), it is an ideal candidate for research on sugarcane. Next-generation sequencing technology has enabled a growing number of studies to focus on genomics. Due to the large amount of omics data available, a centralized platform is necessary for ensuring the consistency, independence and maintainability of these large-scale datasets through storage, analysis and integration. Here, we present a comprehensive database for the E. fulvus genome, EfGD. By using the new high-quality reference genome and its annotations, the EfGD provides the largest whole-genome sequencing reference dataset for E. fulvus, which archives 27 165 protein-coding genes and 55 564 488 SNPs from 202 newly resequenced genomes. Furthermore, we created a user-friendly graphical interface for visualizing genomic diversity, population structure and evolution and provided other tools on an open platform.

Database URL: https://efgenome.ynau.edu.cn

Drug resistance prediction and resistance genes identification in Mycobacterium tuberculosis based on a hierarchical attentive neural network utilizing genome-wide variants

Prediction of antimicrobial resistance based on whole-genome sequencing data has attracted greater attention due to its rapidity and convenience. Numerous machine learning-based studies have used genetic variants to predict drug resistance in Mycobacterium tuberculosis (MTB), assuming that variants are homogeneous, and most of these studies, however, have ignored the essential correlation between variants and corresponding genes when encoding variants, and used a limited number of variants as prediction input. In this study, taking advantage of genome-wide variants for drug-resistance prediction and inspired by natural language processing, we summarize drug resistance prediction into document classification, in which variants are considered as words, mutated genes in an isolate as sentences, and an isolate as a document. We propose a novel hierarchical attentive neural network model (HANN) that helps discover drug resistance-related genes and variants and acquire more interpretable biological results. It captures the interaction among variants in a mutated gene as well as among mutated genes in an isolate. Our results show that for the four first-line drugs of isoniazid (INH), rifampicin (RIF), ethambutol (EMB) and pyrazinamide (PZA), the HANN achieves the optimal area under the ROC curve of 97.90, 99.05, 96.44 and 95.14% and the optimal sensitivity of 94.63, 96.31, 92.56 and 87.05%, respectively. In addition, without any domain knowledge, the model identifies drug resistance-related genes and variants consistent with those confirmed by previous studies, and more importantly, it discovers one more potential drug-resistance-related gene.

MutVis: Automated framework for analysis and visualization of mutational signatures in pathogenic bacterial strains

In recent years, mutational signature analysis has become a routine practice in cancer genomics for classification and diagnosis. Characterizing mutational signatures across species or within genomes of a bacteria helps in understanding their evolution and adaptation. However, an integrated framework for analysis and visualization of mutational signatures in bacterial genome is lacking. Hence, we aim to develop an integrated, automated, open-source and user-friendly framework called MutVis to analyze mutational signatures from bacterial whole genome next generation sequencing data. The current framework integrates various publicly available packages using Snakemake workflow management software, Python and R scripting. MutVis supports variant calling, transition (Ti) and transversion (Tv) graphical representation, generation of mutational count matrix, graphical visualization of base-pair substitution spectrum (BPSs) and mutation signatures extraction. TvTi plots provide the 6 base substitution classification for both genome and gene level. Further resolution of base pair substitution classification is provided as 96-profile BPSs plot. Mutation signatures is derived based on the characteristic pattern observed in BPSs using non-negative matrix factorization. Relative contribution of signatures is given as hierarchically clustered heatmap. This provides information on active signatures in the individual given sample and classify samples according to signature contributions. We demonstrated the MutVis framework using geographically different strains of Mycobacterium tuberculosis, downloaded from PATRIC TB-ARC Antibiotic Resistance Catalog (n = 963). The current framework can be used to study mutation biases and characteristic mutational signatures in bacterial genomes and is freely available at https://github.com/AkshathaPrasanna/MutVis.

Genomic evidence supporting the clonal expansion of extensively drug-resistant tuberculosis bacteria belonging to a rare protoBeijing genotype

Tuberculosis disease (TB), caused by Mycobacterium tuberculosis, is a major public health issue in Thailand. The high prevalence of modern Beijing (Lineage 2.2.1) strains has been associated with multi- and extensively drug-resistant infections (MDR-, XDR-TB), complicating disease control. The impact of rarer proto-Beijing (L2.1) strains is less clear. In our study of thirty-seven L2.1 clinical isolates spanning thirteen years, we found a high prevalence of XDR-TB cases (32.4%). With ≤ 12 pairwise SNP distances, 43.2% of L2.1 patients belong to MDR-TB or XDR-TB transmission clusters suggesting a high level of clonal expansion across four Thai provinces. All XDR-TB (100%) were likely due to transmission rather than inadequate treatment. We found a 47 mutation signature and a partial deletion of the fadD14 gene in the circulating XDR-TB cluster, which can be used for surveillance of this rare and resilient M. tuberculosis strain-type that is causing increasing health burden. We also detected three novel deletion positions, a deletion of 1285 bp within desA3 (Rv3230c), large deletions in the plcB, plcA, and ppe38 gene which may play a role in the virulence, pathogenesis or evolution of the L2.1 strain-type.

Rapid detection of Mycobacterium tuberculosis based on antigen 85B via real-time recombinase polymerase amplification

Tuberculosis (TB), as a common infectious disease, still remains a severe challenge to public health. Due to the unsatisfied clinical needs of currently available diagnostic vehicles, it is desired to establish a new approach for universally detecting Mycobacterium tuberculosis. Herein, we designed a real-time recombinase polymerase amplification (RPA) technology for identifying M. tuberculosis within 20 min at 39°C via custom-designed oligonucleotide primers and probe, which could specifically target antigen 85B (Ag85B). Particularly, the primers F4-R4 produced the fastest fluorescence signal with the probe among four pairs of designed primers in the RPA assays. The optimal primers/probe combination could effectively identify M. tuberculosis with the detection limit of 4·0 copies per μl, as it could not show a positive signal for the genomic DNA from other mycobacteria or pathogens. The Ag85B-based RPA could determine the genomic DNA extracted from M. tuberculosis with high reliability (100%, 22/22). More importantly, when testing clinical sputum samples, the real-time RPA displayed an admirable sensitivity (90%, 95% CI: 80·0-96·0%) and specificity (98%, 95% CI: 89·0-100·0%) compared to traditional smear microscopy, which was similar to the assay of Xpert MTB/RIF. This real-time RPA based Ag85B provides a promising strategy for the rapid and universal diagnosis of TB.