TB-DROP: deep learning-based drug resistance prediction of Mycobacterium tuberculosis utilizing whole genome mutations.

The most widely practiced strategy for constructing the deep learning (DL) prediction model for drug resistance of Mycobacterium tuberculosis (MTB) involves the adoption of ready-made and state-of-the-art architectures usually proposed for non-biological problems. However, the ultimate goal is to construct a customized model for predicting the drug resistance of MTB and eventually for the biological phenotypes based on genotypes. Here, we constructed a DL training framework to standardize and modularize each step during the training process using the latest tensorflow 2 API. A systematic and comprehensive evaluation of each module in the three currently representative models, including Convolutional Neural Network, Denoising Autoencoder, and Wide & Deep, which were adopted by CNNGWP, DeepAMR, and WDNN, respectively, was performed in this framework regarding module contributions in order to assemble a novel model with proper dedicated modules. Based on the whole-genome level mutations, a de novo learning method was developed to overcome the intrinsic limitations of previous models that rely on known drug resistance-associated loci. A customized DL model with the multilayer perceptron architecture was constructed and achieved a competitive performance (the mean sensitivity and specificity were 0.90 and 0.87, respectively) compared to previous ones. The new model developed was applied in an end-to-end user-friendly graphical tool named TB-DROP (TuBerculosis Drug Resistance Optimal Prediction: https://github.com/nottwy/TB-DROP ), in which users only provide sequencing data and TB-DROP will complete analysis within several minutes for one sample. Our study contributes to both a new strategy of model construction and clinical application of deep learning-based drug-resistance prediction methods.

An integrated multi-omics analysis of identifies distinct molecular characteristics in pulmonary infections of Pseudomonas aeruginosa.

Pseudomonas aeruginosa (P. aeruginosa) can cause severe acute infections, including pneumonia and sepsis, and cause chronic infections, commonly in patients with structural respiratory diseases. However, the molecular and pathophysiological mechanisms of P. aeruginosa respiratory infection are largely unknown. Here, we performed assays for transposase-accessible chromatin using sequencing (ATAC-seq), transcriptomics, and quantitative mass spectrometry-based proteomics and ubiquitin-proteomics in P. aeruginosa-infected lung tissues for multi-omics analysis, while ATAC-seq and transcriptomics were also examined in P. aeruginosa-infected mouse macrophages. To identify the pivotal factors that are involved in host immune defense, we integrated chromatin accessibility and gene expression to investigate molecular changes in P. aeruginosa-infected lung tissues combined with proteomics and ubiquitin-proteomics. Our multi-omics investigation discovered a significant concordance for innate immunological and inflammatory responses following P. aeruginosa infection between hosts and alveolar macrophages. Furthermore, we discovered that multi-omics changes in pioneer factors Stat1 and Stat3 play a crucial role in the immunological regulation of P. aeruginosa infection and that their downstream molecules (e.g., Fas) may be implicated in both immunosuppressive and inflammation-promoting processes. Taken together, these findings indicate that transcription factors and their downstream signaling molecules play a critical role in the mobilization and rebalancing of the host immune response against P. aeruginosa infection and may serve as potential targets for bacterial infections and inflammatory diseases, providing insights and resources for omics analyses.

Identification of region of difference and H37Rv-related deletion in Mycobacterium tuberculosis complex by structural variant detection and genome assembly.

Mycobacterium tuberculosis complex (MTBC), the main cause of TB in humans and animals, is an extreme example of genetic homogeneity, whereas it is still nevertheless separated into various lineages by numerous typing methods, which differ in phenotype, virulence, geographic distribution, and host preference. The large sequence polymorphism (LSP), incorporating region of difference (RD) and H37Rv-related deletion (RvD), is considered to be a powerful means of constructing phylogenetic relationships within MTBC. Although there have been many studies on LSP already, focusing on the distribution of RDs in MTBC and their impact on MTB phenotypes, a crumb of new lineages or sub-lineages have been excluded and RvDs have received less attention. We, therefore, sampled a dataset of 1,495 strains, containing 113 lineages from the laboratory collection, to screen for RDs and RvDs by structural variant detection and genome assembly, and examined the distribution of RvDs in MTBC, including RvD2, RvD5, and cobF region. Consistent with genealogical delineation by single nucleotide polymorphism (SNP), we identified 125 RDs and 5 RvDs at the species, lineage, or sub-lineage levels. The specificities of RDs and RvDs were further investigated in the remaining 10,218 strains, suggesting that most of them were highly specific to distinct phylogenetic groups, could be used as stable genetic markers in genotyping. More importantly, we identified 34 new lineage or evolutionary branch specific RDs and 2 RvDs, also demonstrated the distribution of known RDs and RvDs in MTBC. This study provides novel details about deletion events that have occurred in distinct phylogenetic groups and may help to understand the genealogical differentiation.

Identification of Mobile Colistin Resistance Gene mcr-10 in Disinfectant and Antibiotic Resistant Escherichia coli from Disinfected Tableware.

The widespread escalation of bacterial resistance threatens the safety of the food chain. To investigate the resistance characteristics of E. coli strains isolated from disinfected tableware against both disinfectants and antibiotics, 311 disinfected tableware samples, including 54 chopsticks, 32 dinner plates, 61 bowls, 11 cups, and three spoons were collected in Chengdu, Sichuan Province, China to screen for disinfectant- (benzalkonium chloride and cetylpyridinium chloride) and tigecycline-resistant isolates, which were then subjected to antimicrobial susceptibility testing and whole genome sequencing (WGS). The coliform-positive detection rate was 51.8% (161/311) and among 161 coliform-positive samples, eight E. coli strains were multidrug-resistant to benzalkonium chloride, cetylpyridinium chloride, ampicillin, and tigecycline. Notably, a recently described mobile colistin resistance gene mcr-10 present on the novel IncFIB-type plasmid of E. coli EC2641 screened was able to successfully transform the resistance. Global phylogenetic analysis revealed E. coli EC2641 clustered together with two clinically disinfectant- and colistin-multidrug-resistant E. coli strains from the US. This is the first report of mcr-10-bearing E. coli detected in disinfected tableware, suggesting that continuous monitoring of resistance genes in the catering industry is essential to understand and respond to the transmission of antibiotic resistance genes from the environment and food to humans and clinics.

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.