A White-Box Machine Learning Approach for Revealing Antibiotic Mechanisms of Action

Metagenomics reveals the resistance patterns of electrochemically treated erythromycin fermentation residue

Erythromycin fermentation residue (EFR) represents a typical hazardous waste produced by the microbial pharmaceutical industry. Although electrolysis is promising for EFR disposal, its microbial threats remain unclear. Herein, metagenomics was coupled with the random forest technique to decipher the antibiotic resistance patterns of electrochemically treated EFR. Results showed that 95.75% of erythromycin could be removed in 2 hr. Electrolysis temporarily influenced EFR microbiota, where the relative abundances of Proteobacteria and Actinobacteria increased, while those of Fusobacteria, Firmicutes, and Bacteroidetes decreased. A total of 505 antibiotic resistance gene (ARG) subtypes encoding resistance to 21 antibiotic types and 150 mobile genetic elements (MGEs), mainly including plasmid (72) and transposase (52) were assembled in EFR. Significant linear regression models were identified among microbial richness, ARG subtypes, and MGE numbers (r2=0.50-0.81, p< 0.001). Physicochemical factors of EFR (Total nitrogen, total organic carbon, protein, and humus) regulated ARG and MGE assembly (%IncMSE value = 5.14-14.85). The core ARG, MGE, and microbe sets (93.08%-99.85%) successfully explained 89.71%-92.92% of total ARG and MGE abundances. Specifically, gene aph(3')-I, transposase tnpA, and Mycolicibacterium were the primary drivers of the resistance dissemination system. This study also proposes efficient resistance mitigation measures, and provides recommendations for future management of antibiotic fermentation residue.

Glucocorticoid therapy for sepsis in the AI era: a survey on current and future approaches

Sepsis, a life-threatening medical condition, manifests as new or worsening organ failures due to a dysregulated host response to infection. Many patients with sepsis have manifested a hyperinflammatory phenotype leading to the identification of inflammatory modulation by corticosteroids as a key treatment modality. However, the optimal use of corticosteroids in sepsis treatment remains a contentious subject, necessitating a deeper understanding of their physiological and pharmacological effects. Our study conducts a comprehensive review of randomized controlled trials (RCTs) focusing on traditional corticosteroid treatment in sepsis, alongside an analysis of evolving clinical guidelines. Additionally, we explore the emerging role of artificial intelligence (AI) in medicine, particularly in diagnosing, prognosticating, and treating sepsis. AI's advanced data processing capabilities reveal new avenues for enhancing corticosteroid therapeutic strategies in sepsis. The integration of AI in sepsis treatment has the potential to address existing gaps in knowledge, especially in the application of corticosteroids. Our findings suggest that combining corticosteroid therapy with AI-driven insights could lead to more personalized and effective sepsis treatments. This approach holds promise for improving clinical outcomes and presents a significant advancement in the management of this complex and often fatal condition.

Antibiotic tolerance among clinical isolates: mechanisms, detection, prevalence, and significance

SUMMARYAntibiotic treatment failures in the absence of resistance are not uncommon. Recently, attention has grown around the phenomenon of antibiotic tolerance, an underappreciated contributor to recalcitrant infections first detected in the 1970s. Tolerance describes the ability of a bacterial population to survive transient exposure to an otherwise lethal concentration of antibiotic without exhibiting resistance. With advances in genomics, we are gaining a better understanding of the molecular mechanisms behind tolerance, and several studies have sought to examine the clinical prevalence of tolerance. Attempts have also been made to assess the clinical significance of tolerance through in vivo infection models and prospective/retrospective clinical studies. Here, we review the data available on the molecular mechanisms, detection, prevalence, and clinical significance of genotypic tolerance that span ~50 years. We discuss the need for standardized methodology and interpretation criteria for tolerance detection and the impact that methodological inconsistencies have on our ability to accurately assess the scale of the problem. In terms of the clinical significance of tolerance, studies suggest that tolerance contributes to worse outcomes for patients (e.g., higher mortality, prolonged hospitalization), but historical data from animal models are varied. Furthermore, we lack the necessary information to effectively treat tolerant infections. Overall, while the tolerance field is gaining much-needed traction, the underlying clinical significance of tolerance that underpins all tolerance research is still far from clear and requires attention.

Predicting bacterial fitness in Mycobacterium tuberculosis with transcriptional regulatory network-informed interpretable machine learning

Mycobacterium tuberculosis (Mtb) is the causative agent of tuberculosis disease, the greatest source of global mortality by a bacterial pathogen. Mtb adapts and responds to diverse stresses such as antibiotics by inducing transcriptional stress-response regulatory programs. Understanding how and when these mycobacterial regulatory programs are activated could enable novel treatment strategies for potentiating the efficacy of new and existing drugs. Here we sought to define and analyze Mtb regulatory programs that modulate bacterial fitness. We assembled a large Mtb RNA expression compendium and applied these to infer a comprehensive Mtb transcriptional regulatory network and compute condition-specific transcription factor activity profiles. We utilized transcriptomic and functional genomics data to train an interpretable machine learning model that can predict Mtb fitness from transcription factor activity profiles. We demonstrated that this transcription factor activity-based model can successfully predict Mtb growth arrest and growth resumption under hypoxia and reaeration using only RNA-seq expression data as a starting point. These integrative network modeling and machine learning analyses thus enable the prediction of mycobacterial fitness under different environmental and genetic contexts. We envision these models can potentially inform the future design of prognostic assays and therapeutic intervention that can cripple Mtb growth and survival to cure tuberculosis disease.

Anticancer agent 5-fluorouracil reverses meropenem resistance in carbapenem-resistant Gram-negative pathogens

The global increasing incidence of clinical infections caused by carbapenem-resistant Gram-negative pathogens requires urgent and effective treatment strategies. Antibiotic adjuvants represent a promising approach to enhance the efficacy of meropenem against carbapenem-resistant bacteria. This study shows that the anticancer agent 5-fluorouracil (5-FU, 50 µM) significantly reduced the minimum inhibitory concentration of meropenem against blaNDM-5 positive Escherichia coli by 32-fold through cell-based high-throughput screening. Further pharmacological studies indicated that 5-FU exhibited potentiation effects on carbapenem antibiotics against 42 Gram-negative bacteria producing either metallo-β-lactamases (MBLs), such as NDM and IMP, or serine β-lactamases (Ser-BLs), like KPC and OXA. These bacteria included E. coli, Klebsiella pneumoniae, Pseudomonas aeruginosa and Acinetobacter spp., 32 of which were obtained from human clinical samples. Mechanistic investigations revealed that 5-FU inhibited the transcription and expression of the blaNDM-5 gene. In addition, 5-FU combined with meropenem enhanced bacterial metabolism, and stimulated the production of reactive oxygen species (ROS), thereby rendering bacteria more susceptible to meropenem. In a mouse systemic infection model, 5-FU combined with meropenem reduced bacterial loads and effectively elevated the survival rate of 83.3%, compared with 16.7% with meropenem monotherapy. Collectively, these findings indicate the potential of 5-FU as a novel meropenem adjuvant to improve treatment outcomes against infections caused by carbapenem-resistant bacteria.

Cell factory design with advanced metabolic modelling empowered by artificial intelligence

Advances in synthetic biology and artificial intelligence (AI) have provided new opportunities for modern biotechnology. High-performance cell factories, the backbone of industrial biotechnology, are ultimately responsible for determining whether a bio-based product succeeds or fails in the fierce competition with petroleum-based products. To date, one of the greatest challenges in synthetic biology is the creation of high-performance cell factories in a consistent and efficient manner. As so-called white-box models, numerous metabolic network models have been developed and used in computational strain design. Moreover, great progress has been made in AI-powered strain engineering in recent years. Both approaches have advantages and disadvantages. Therefore, the deep integration of AI with metabolic models is crucial for the construction of superior cell factories with higher titres, yields and production rates. The detailed applications of the latest advanced metabolic models and AI in computational strain design are summarized in this review. Additionally, approaches for the deep integration of AI and metabolic models are discussed. It is anticipated that advanced mechanistic metabolic models powered by AI will pave the way for the efficient construction of powerful industrial chassis strains in the coming years.

Machine learning for the advancement of genome-scale metabolic modeling

Constraint-based modeling (CBM) has evolved as the core systems biology tool to map the interrelations between genotype, phenotype, and external environment. The recent advancement of high-throughput experimental approaches and multi-omics strategies has generated a plethora of new and precise information from wide-ranging biological domains. On the other hand, the continuously growing field of machine learning (ML) and its specialized branch of deep learning (DL) provide essential computational architectures for decoding complex and heterogeneous biological data. In recent years, both multi-omics and ML have assisted in the escalation of CBM. Condition-specific omics data, such as transcriptomics and proteomics, helped contextualize the model prediction while analyzing a particular phenotypic signature. At the same time, the advanced ML tools have eased the model reconstruction and analysis to increase the accuracy and prediction power. However, the development of these multi-disciplinary methodological frameworks mainly occurs independently, which limits the concatenation of biological knowledge from different domains. Hence, we have reviewed the potential of integrating multi-disciplinary tools and strategies from various fields, such as synthetic biology, CBM, omics, and ML, to explore the biochemical phenomenon beyond the conventional biological dogma. How the integrative knowledge of these intersected domains has improved bioengineering and biomedical applications has also been highlighted. We categorically explained the conventional genome-scale metabolic model (GEM) reconstruction tools and their improvement strategies through ML paradigms. Further, the crucial role of ML and DL in omics data restructuring for GEM development has also been briefly discussed. Finally, the case-study-based assessment of the state-of-the-art method for improving biomedical and metabolic engineering strategies has been elaborated. Therefore, this review demonstrates how integrating experimental and in silico strategies can help map the ever-expanding knowledge of biological systems driven by condition-specific cellular information. This multiview approach will elevate the application of ML-based CBM in the biomedical and bioengineering fields for the betterment of society and the environment.

Systems modeling of oncogenic G-protein and GPCR signaling reveals unexpected differences in downstream pathway activation

Mathematical models of biochemical reaction networks are an important and emerging tool for the study of cell signaling networks involved in disease processes. One promising potential application of such mathematical models is the study of how disease-causing mutations promote the signaling phenotype that contributes to the disease. It is commonly assumed that one must have a thorough characterization of the network readily available for mathematical modeling to be useful, but we hypothesized that mathematical modeling could be useful when there is incomplete knowledge and that it could be a tool for discovery that opens new areas for further exploration. In the present study, we first develop a mechanistic mathematical model of a G-protein coupled receptor signaling network that is mutated in almost all cases of uveal melanoma and use model-driven explorations to uncover and explore multiple new areas for investigating this disease. Modeling the two major, mutually-exclusive, oncogenic mutations (Gαq/11 and CysLT2R) revealed the potential for previously unknown qualitative differences between seemingly interchangeable disease-promoting mutations, and our experiments confirmed oncogenic CysLT2R was impaired at activating the FAK/YAP/TAZ pathway relative to Gαq/11. This led us to hypothesize that CYSLTR2 mutations in UM must co-occur with other mutations to activate FAK/YAP/TAZ signaling, and our bioinformatic analysis uncovers a role for co-occurring mutations involving the plexin/semaphorin pathway, which has been shown capable of activating this pathway. Overall, this work highlights the power of mechanism-based computational systems biology as a discovery tool that can leverage available information to open new research areas.

Bioenergetic stress potentiates antimicrobial resistance and persistence

Antimicrobial resistance (AMR) is a global health crisis and there is an urgent need to better understand AMR mechanisms. Antibiotic treatment alters several aspects of bacterial physiology, including increased ATP utilization, carbon metabolism, and reactive oxygen species (ROS) formation. However, how the "bioenergetic stress" induced by increased ATP utilization affects treatment outcomes is unknown. Here we utilized a synthetic biology approach to study the direct effects of bioenergetic stress on antibiotic efficacy. We engineered a genetic system that constitutively hydrolyzes ATP or NADH in Escherichia coli. We found that bioenergetic stress potentiates AMR evolution via enhanced ROS production, mutagenic break repair, and transcription-coupled repair. We also find that bioenergetic stress potentiates antimicrobial persistence via potentiated stringent response activation. We propose a unifying model that antibiotic-induced antimicrobial resistance and persistence is caused by antibiotic-induced. This has important implications for preventing or curbing the spread of AMR infections.

Metabolomics analysis of the lactobacillus plantarum ATCC 14917 response to antibiotic stress

BACKGROUND

Lactobacillus plantarum has been found to play a significant role in maintaining the balance of intestinal flora in the human gut. However, it is sensitive to commonly used antibiotics and is often incidentally killed during treatment. We attempted to identify a means to protect L. plantarum ATCC14917 from the metabolic changes caused by two commonly used antibiotics, ampicillin, and doxycycline. We examined the metabolic changes under ampicillin and doxycycline treatment and assessed the protective effects of adding key exogenous metabolites.

RESULTS

Using metabolomics, we found that under the stress of ampicillin or doxycycline, L. plantarum ATCC14917 exhibited reduced metabolic activity, with purine metabolism a key metabolic pathway involved in this change. We then screened the key biomarkers in this metabolic pathway, guanine and adenosine diphosphate (ADP). The exogenous addition of each of these two metabolites significantly reduced the lethality of ampicillin and doxycycline on L. plantarum ATCC14917. Because purine metabolism is closely related to the production of reactive oxygen species (ROS), the results showed that the addition of guanine or ADP reduced intracellular ROS levels in L. plantarum ATCC14917. Moreover, the killing effects of ampicillin and doxycycline on L. plantarum ATCC14917 were restored by the addition of a ROS accelerator in the presence of guanine or ADP.

CONCLUSIONS

The metabolic changes of L. plantarum ATCC14917 under antibiotic treatments were determined. Moreover, the metabolome information that was elucidated can be used to help L. plantarum cope with adverse stress, which will help probiotics become less vulnerable to antibiotics during clinical treatment.

Deciphering the Antibacterial Mechanisms of 5-Fluorouracil in Escherichia coli through Biochemical and Transcriptomic Analyses

The emergence of carbapenem-resistant Gram-negative pathogens presents a clinical challenge in infection treatment, prompting the repurposing of existing drugs as an essential strategy to address this crisis. Although the anticancer drug 5-fluorouracil (5-FU) has been recognized for its antibacterial properties, its mechanisms are not fully understood. Here, we found that the minimal inhibitory concentration (MIC) of 5-FU against Escherichia coli was 32-64 µg/mL, including strains carrying blaNDM-5, which confers resistance to carbapenems. We further elucidated the antibacterial mechanism of 5-FU against E. coli by using genetic and biochemical analyses. We revealed that the mutation of uracil phosphoribosyltransferase-encoding gene upp increased the MIC of 5-FU against E. coli by 32-fold, indicating the role of the upp gene in 5-FU resistance. Additionally, transcriptomic analysis of E. coli treated with 5-FU at 8 µg/mL and 32 µg/mL identified 602 and 1082 differentially expressed genes involved in carbon and nucleic acid metabolism, DNA replication, and repair pathways. The biochemical assays showed that 5-FU induced bacterial DNA damage, significantly increased intracellular ATP levels and the NAD+/NADH ratio, and promoted reactive oxygen species (ROS) production. These findings suggested that 5-FU may exert antibacterial effects on E. coli through multiple pathways, laying the groundwork for its further development as a therapeutic candidate against carbapenem-resistant bacterial infections.

Providing insight into the mechanism of action of cationic lipidated oligomers using metabolomics

The increasing resistance of clinically relevant microbes against current commercially available antimicrobials underpins the urgent need for alternative and novel treatment strategies. Cationic lipidated oligomers (CLOs) are innovative alternatives to antimicrobial peptides and have reported antimicrobial potential. An understanding of their antimicrobial mechanism of action is required to rationally design future treatment strategies for CLOs, either in monotherapy or synergistic combinations. In the present study, metabolomics was used to investigate the potential metabolic pathways involved in the mechanisms of antibacterial activity of one CLO, C12-o-(BG-D)-10, which we have previously shown to be effective against methicillin-resistant Staphylococcus aureus (MRSA) ATCC 43300. The metabolomes of MRSA ATCC 43300 at 1, 3, and 6 h following treatment with C12-o-(BG-D)-10 (48 µg/mL, i.e., 3× MIC) were compared to those of the untreated controls. Our findings reveal that the studied CLO, C12-o-(BG-D)-10, disorganized the bacterial membrane as the first step toward its antimicrobial effect, as evidenced by marked perturbations in the bacterial membrane lipids and peptidoglycan biosynthesis observed at early time points, i.e., 1 and 3 h. Central carbon metabolism and the biosynthesis of DNA, RNA, and arginine were also vigorously perturbed, mainly at early time points. Moreover, bacterial cells were under osmotic and oxidative stress across all time points, as evident by perturbations of trehalose biosynthesis and pentose phosphate shunt. Overall, this metabolomics study has, for the first time, revealed that the antimicrobial action of C12-o-(BG-D)-10 may potentially stem from the dysregulation of multiple metabolic pathways.IMPORTANCEAntimicrobial resistance poses a significant challenge to healthcare systems worldwide. Novel anti-infective therapeutics are urgently needed to combat drug-resistant microorganisms. Cationic lipidated oligomers (CLOs) show promise as new antibacterial agents against Gram-positive pathogens like methicillin-resistant Staphylococcus aureus (MRSA). Understanding their molecular mechanism(s) of antimicrobial action may help design synergistic CLO treatments along with monotherapy. Here, we describe the first metabolomics study to investigate the killing mechanism(s) of CLOs against MRSA. The results of our study indicate that the CLO, C12-o-(BG-D)-10, had a notable impact on the biosynthesis and organization of the bacterial cell envelope. C12-o-(BG-D)-10 also inhibits arginine, histidine, central carbon metabolism, and trehalose production, adding to its antibacterial characteristics. This work illuminates the unique mechanism of action of C12-o-(BG-D)-10 and opens an avenue to design innovative antibacterial oligomers/polymers for future clinical applications.

Progress of Antimicrobial Mechanisms of Stilbenoids

Antimicrobial drugs have made outstanding contributions to the treatment of pathogenic infections. However, the emergence of drug resistance continues to be a major threat to human health in recent years, and therefore, the search for novel antimicrobial drugs is particularly urgent. With a deeper understanding of microbial habits and drug resistance mechanisms, various creative strategies for the development of novel antibiotics have been proposed. Stilbenoids, characterized by a C6-C2-C6 carbon skeleton, have recently been widely recognized for their flexible antimicrobial roles. Here, we comprehensively summarize the mode of action of stilbenoids from the viewpoint of their direct antimicrobial properties, antibiofilm and antivirulence activities and their role in reversing drug resistance. This review will provide an important reference for the future development and research into the mechanisms of stilbenoids as antimicrobial agents.

Prediction of anticancer drug sensitivity using an interpretable model guided by deep learning

BACKGROUND

The prediction of drug sensitivity plays a crucial role in improving the therapeutic effect of drugs. However, testing the effectiveness of drugs is challenging due to the complex mechanism of drug reactions and the lack of interpretability in most machine learning and deep learning methods. Therefore, it is imperative to establish an interpretable model that receives various cell line and drug feature data to learn drug response mechanisms and achieve stable predictions between available datasets.

RESULTS

This study proposes a new and interpretable deep learning model, DrugGene, which integrates gene expression, gene mutation, gene copy number variation of cancer cells, and chemical characteristics of anticancer drugs to predict their sensitivity. This model comprises two different branches of neural networks, where the first involves a hierarchical structure of biological subsystems that uses the biological processes of human cells to form a visual neural network (VNN) and an interpretable deep neural network for human cancer cells. DrugGene receives genotype input from the cell line and detects changes in the subsystem states. We also employ a traditional artificial neural network (ANN) to capture the chemical structural features of drugs. DrugGene generates final drug response predictions by combining VNN and ANN and integrating their outputs into a fully connected layer. The experimental results using drug sensitivity data extracted from the Cancer Drug Sensitivity Genome Database and the Cancer Treatment Response Portal v2 reveal that the proposed model is better than existing prediction methods. Therefore, our model achieves higher accuracy, learns the reaction mechanisms between anticancer drugs and cell lines from various features, and interprets the model's predicted results.

CONCLUSIONS

Our method utilizes biological pathways to construct neural networks, which can use genotypes to monitor changes in the state of network subsystems, thereby interpreting the prediction results in the model and achieving satisfactory prediction accuracy. This will help explore new directions in cancer treatment. More available code resources can be downloaded for free from GitHub ( https://github.com/pangweixiong/DrugGene ).

Purine and pyrimidine synthesis differently affect the strength of the inoculum effect for aminoglycoside and β-lactam antibiotics

The inoculum effect has been observed for nearly all antibiotics and bacterial species. However, explanations accounting for its occurrence and strength are lacking. We previously found that growth productivity, which captures the relationship between [ATP] and growth, can account for the strength of the inoculum effect for bactericidal antibiotics. However, the molecular pathway(s) underlying this relationship, and therefore determining the inoculum effect, remain undiscovered. We show that nucleotide synthesis can determine the relationship between [ATP] and growth, and thus the strength of inoculum effect in an antibiotic class-dependent manner. Specifically, and separate from activity through the tricarboxylic acid cycle, we find that transcriptional activity of genes involved in purine and pyrimidine synthesis can predict the strength of the inoculum effect for β-lactam and aminoglycosides antibiotics, respectively. Our work highlights the antibiotic class-specific effect of purine and pyrimidine synthesis on the severity of the inoculum effect and paves the way for intervention strategies to reduce the inoculum effect in the clinic.

Identifying and mathematically modeling the time-course of extracellular metabolic markers associated with resistance to ceftolozane/tazobactam in Pseudomonas aeruginosa

Extracellular bacterial metabolites have potential as markers of bacterial growth and resistance emergence but have not been evaluated in dynamic in vitro studies. We investigated the dynamic metabolomic footprint of a multidrug-resistant hypermutable Pseudomonas aeruginosa isolate exposed to ceftolozane/tazobactam as continuous infusion (4.5 g/day, 9 g/day) in a hollow-fiber infection model over 7-9 days in biological replicates (n = 5). Bacterial samples were collected at 0, 7, 23, 47, 71, 95, 143, 167, 191, and 215 h, the supernatant quenched, and extracellular metabolites extracted. Metabolites were analyzed via untargeted metabolomics, including hierarchical clustering and correlation with quantified total and resistant bacterial populations. The time-courses of five (of 1,921 detected) metabolites from enriched pathways were mathematically modeled. Absorbed L-arginine and secreted L-ornithine were highly correlated with the total bacterial population (r -0.79 and 0.82, respectively, P<0.0001). Ribose-5-phosphate, sedoheptulose-7-phosphate, and trehalose-6-phosphate correlated with the resistant subpopulation (0.64, 0.64, and 0.67, respectively, P<0.0001) and were likely secreted due to resistant growth overcoming oxidative and osmotic stress induced by ceftolozane/tazobactam. Using pharmacokinetic/pharmacodynamic-based transduction models, these metabolites were successfully modeled based on the total or resistant bacterial populations. The models well described the abundance of each metabolite across the differing time-course profiles of biological replicates, based on bacterial killing and, importantly, resistant regrowth. These proof-of-concept studies suggest that further exploration is warranted to determine the generalizability of these findings. The metabolites modeled here are not exclusive to bacteria. Future studies may use this approach to identify bacteria-specific metabolites correlating with resistance, which would ultimately be extremely useful for clinical translation.

Machine learning identifies key metabolic reactions in bacterial growth on different carbon sources

Carbon source-dependent control of bacterial growth is fundamental to bacterial physiology and survival. However, pinpointing the metabolic steps important for cell growth is challenging due to the complexity of cellular networks. Here, the elastic net model and multilayer perception model that integrated genome-wide gene-deletion data and simulated flux distributions were constructed to identify metabolic reactions beneficial or detrimental to Escherichia coli grown on 30 different carbon sources. Both models outperformed traditional in silico methods by identifying not just essential reactions but also nonessential ones that promote growth. They successfully predicted metabolic reactions beneficial to cell growth, with high convergence between the models. The models revealed that biosynthetic pathways generally promote growth across various carbon sources, whereas the impact of energy-generating pathways varies with the carbon source. Intriguing predictions were experimentally validated for findings beyond experimental training data and the impact of various carbon sources on the glyoxylate shunt, pyruvate dehydrogenase reaction, and redundant purine biosynthesis reactions. These highlight the practical significance and predictive power of the models for understanding and engineering microbial metabolism.

Towards a hybrid model-driven platform based on flux balance analysis and a machine learning pipeline for biosystem design

Metabolic modeling and machine learning (ML) are crucial components of the evolving next-generation tools in systems and synthetic biology, aiming to unravel the intricate relationship between genotype, phenotype, and the environment. Nonetheless, the comprehensive exploration of integrating these two frameworks, and fully harnessing the potential of fluxomic data, remains an unexplored territory. In this study, we present, rigorously evaluate, and compare ML-based techniques for data integration. The hybrid model revealed that the overexpression of six target genes and the knockout of seven target genes contribute to enhanced ethanol production. Specifically, we investigated the influence of succinate dehydrogenase (SDH) on ethanol biosynthesis in Saccharomyces cerevisiae through shake flask experiments. The findings indicate a noticeable increase in ethanol yield, ranging from 6 % to 10 %, in SDH subunit gene knockout strains compared to the wild-type strain. Moreover, in pursuit of a high-yielding strain for ethanol production, dual-gene deletion experiments were conducted targeting glycerol-3-phosphate dehydrogenase (GPD) and SDH. The results unequivocally demonstrate significant enhancements in ethanol production for the engineered strains Δsdh4Δgpd1, Δsdh5Δgpd1, Δsdh6Δgpd1, Δsdh4Δgpd2, Δsdh5Δgpd2, and Δsdh6Δgpd2, with improvements of 21.6 %, 27.9 %, and 22.7 %, respectively. Overall, the results highlighted that integrating mechanistic flux features substantially improves the prediction of gene knockout strains not accounted for in metabolic reconstructions. In addition, the finding in this study delivers valuable tools for comprehending and manipulating intricate phenotypes, thereby enhancing prediction accuracy and facilitating deeper insights into mechanistic aspects within the field of synthetic biology.

Optimising machine learning prediction of minimum inhibitory concentrations in Klebsiella pneumoniae

Minimum Inhibitory Concentrations (MICs) are the gold standard for quantitatively measuring antibiotic resistance. However, lab-based MIC determination can be time-consuming and suffers from low reproducibility, and interpretation as sensitive or resistant relies on guidelines which change over time. Genome sequencing and machine learning promise to allow in silico MIC prediction as an alternative approach which overcomes some of these difficulties, albeit the interpretation of MIC is still needed. Nevertheless, precisely how we should handle MIC data when dealing with predictive models remains unclear, since they are measured semi-quantitatively, with varying resolution, and are typically also left- and right-censored within varying ranges. We therefore investigated genome-based prediction of MICs in the pathogen Klebsiella pneumoniae using 4367 genomes with both simulated semi-quantitative traits and real MICs. As we were focused on clinical interpretation, we used interpretable rather than black-box machine learning models, namely, Elastic Net, Random Forests, and linear mixed models. Simulated traits were generated accounting for oligogenic, polygenic, and homoplastic genetic effects with different levels of heritability. Then we assessed how model prediction accuracy was affected when MICs were framed as regression and classification. Our results showed that treating the MICs differently depending on the number of concentration levels of antibiotic available was the most promising learning strategy. Specifically, to optimise both prediction accuracy and inference of the correct causal variants, we recommend considering the MICs as continuous and framing the learning problem as a regression when the number of observed antibiotic concentration levels is large, whereas with a smaller number of concentration levels they should be treated as a categorical variable and the learning problem should be framed as a classification. Our findings also underline how predictive models can be improved when prior biological knowledge is taken into account, due to the varying genetic architecture of each antibiotic resistance trait. Finally, we emphasise that incrementing the population database is pivotal for the future clinical implementation of these models to support routine machine-learning based diagnostics.

Logic-based mechanistic machine learning on high-content images reveals how drugs differentially regulate cardiac fibroblasts

Fibroblasts are essential regulators of extracellular matrix deposition following cardiac injury. These cells exhibit highly plastic responses in phenotype during fibrosis in response to environmental stimuli. Here, we test whether and how candidate anti-fibrotic drugs differentially regulate measures of cardiac fibroblast phenotype, which may help identify treatments for cardiac fibrosis. We conducted a high-content microscopy screen of human cardiac fibroblasts treated with 13 clinically relevant drugs in the context of TGFβ and/or IL-1β, measuring phenotype across 137 single-cell features. We used the phenotypic data from our high-content imaging to train a logic-based mechanistic machine learning model (LogiMML) for fibroblast signaling. The model predicted how pirfenidone and Src inhibitor WH-4-023 reduce actin filament assembly and actin-myosin stress fiber formation, respectively. Validating the LogiMML model prediction that PI3K partially mediates the effects of Src inhibition, we found that PI3K inhibition reduces actin-myosin stress fiber formation and procollagen I production in human cardiac fibroblasts. In this study, we establish a modeling approach combining the strengths of logic-based network models and regularized regression models. We apply this approach to predict mechanisms that mediate the differential effects of drugs on fibroblasts, revealing Src inhibition acting via PI3K as a potential therapy for cardiac fibrosis.

Antimicrobial resistance crisis: could artificial intelligence be the solution?

Antimicrobial resistance is a global public health threat, and the World Health Organization (WHO) has announced a priority list of the most threatening pathogens against which novel antibiotics need to be developed. The discovery and introduction of novel antibiotics are time-consuming and expensive. According to WHO's report of antibacterial agents in clinical development, only 18 novel antibiotics have been approved since 2014. Therefore, novel antibiotics are critically needed. Artificial intelligence (AI) has been rapidly applied to drug development since its recent technical breakthrough and has dramatically improved the efficiency of the discovery of novel antibiotics. Here, we first summarized recently marketed novel antibiotics, and antibiotic candidates in clinical development. In addition, we systematically reviewed the involvement of AI in antibacterial drug development and utilization, including small molecules, antimicrobial peptides, phage therapy, essential oils, as well as resistance mechanism prediction, and antibiotic stewardship.

The protective role of commensal gut microbes and their metabolites against bacterial pathogens

Multidrug-resistant microorganisms have become a major public health concern around the world. The gut microbiome is a gold mine for bioactive compounds that protect the human body from pathogens. We used a multi-omics approach that integrated whole-genome sequencing (WGS) of 74 commensal gut microbiome isolates with metabolome analysis to discover their metabolic interaction with Salmonella and other antibiotic-resistant pathogens. We evaluated differences in the functional potential of these selected isolates based on WGS annotation profiles. Furthermore, the top altered metabolites in co-culture supernatants of selected commensal gut microbiome isolates were identified including a series of dipeptides and examined for their ability to prevent the growth of various antibiotic-resistant bacteria. Our results provide compelling evidence that the gut microbiome produces metabolites, including the compound class of dipeptides that can potentially be applied for anti-infection medication, especially against antibiotic-resistant pathogens. Our established pipeline for the discovery and validation of bioactive metabolites from the gut microbiome as novel candidates for multidrug-resistant infections represents a new avenue for the discovery of antimicrobial lead structures.

Cancer immunotherapy efficacy and machine learning

INTRODUCTION

Immunotherapy is one of the major breakthroughs in the treatment of cancer, and it has become a powerful clinical strategy, however, not all patients respond to immune checkpoint blockade and other immunotherapy strategies. Applying machine learning (ML) techniques to predict the efficacy of cancer immunotherapy is useful for clinical decision-making.

AREAS COVERED

Applying ML including deep learning (DL) in radiomics, pathomics, tumor microenvironment (TME) and immune-related genes analysis to predict immunotherapy efficacy. The studies in this review were searched from PubMed and ClinicalTrials.gov (January 2023).

EXPERT OPINION

An increasing number of studies indicate that ML has been applied to various aspects of oncology research, with the potential to provide more effective individualized immunotherapy strategies and enhance treatment decisions. With advances in ML technology, more efficient methods of predicting the efficacy of immunotherapy may become available in the future.

Logic-based mechanistic machine learning on high-content images reveals how drugs differentially regulate cardiac fibroblasts

Fibroblasts are essential regulators of extracellular matrix deposition following cardiac injury. These cells exhibit highly plastic responses in phenotype during fibrosis in response to environmental stimuli. Here, we test whether and how candidate anti-fibrotic drugs differentially regulate measures of cardiac fibroblast phenotype, which may help identify treatments for cardiac fibrosis. We conducted a high content microscopy screen of human cardiac fibroblasts treated with 13 clinically relevant drugs in the context of TGFβ and/or IL-1β, measuring phenotype across 137 single-cell features. We used the phenotypic data from our high content imaging to train a logic-based mechanistic machine learning model (LogiMML) for fibroblast signaling. The model predicted how pirfenidone and Src inhibitor WH-4-023 reduce actin filament assembly and actin-myosin stress fiber formation, respectively. Validating the LogiMML model prediction that PI3K partially mediates the effects of Src inhibition, we found that PI3K inhibition reduces actin-myosin stress fiber formation and procollagen I production in human cardiac fibroblasts. In this study, we establish a modeling approach combining the strengths of logic-based network models and regularized regression models, apply this approach to predict mechanisms that mediate the differential effects of drugs on fibroblasts, revealing Src inhibition acting via PI3K as a potential therapy for cardiac fibrosis.

The Combination of Antibiotic and Non-Antibiotic Compounds Improves Antibiotic Efficacy against Multidrug-Resistant Bacteria

Bacterial antibiotic resistance, especially the emergence of multidrug-resistant (MDR) strains, urgently requires the development of effective treatment strategies. It is always of interest to delve into the mechanisms of resistance to current antibiotics and target them to promote the efficacy of existing antibiotics. In recent years, non-antibiotic compounds have played an important auxiliary role in improving the efficacy of antibiotics and promoting the treatment of drug-resistant bacteria. The combination of non-antibiotic compounds with antibiotics is considered a promising strategy against MDR bacteria. In this review, we first briefly summarize the main resistance mechanisms of current antibiotics. In addition, we propose several strategies to enhance antibiotic action based on resistance mechanisms. Then, the research progress of non-antibiotic compounds that can promote antibiotic-resistant bacteria through different mechanisms in recent years is also summarized. Finally, the development prospects and challenges of these non-antibiotic compounds in combination with antibiotics are discussed.

KatG catalase deficiency confers bedaquiline hyper-susceptibility to isoniazid resistant Mycobacterium tuberculosis

Multidrug-resistant tuberculosis (MDR-TB) is a growing source of global mortality and threatens global control of tuberculosis (TB) disease. The diarylquinoline bedaquiline (BDQ) recently emerged as a highly efficacious drug against MDR-TB, defined as resistance to the first-line drugs isoniazid (INH) and rifampin. INH resistance is primarily caused by loss-of-function mutations in the catalase KatG, but mechanisms underlying BDQ's efficacy against MDR-TB remain unknown. Here we employ a systems biology approach to investigate BDQ hyper-susceptibility in INH-resistant Mycobacterium tuberculosis . We found hyper-susceptibility to BDQ in INH-resistant cells is due to several physiological changes induced by KatG deficiency, including increased susceptibility to reactive oxygen species and DNA damage, remodeling of transcriptional programs, and metabolic repression of folate biosynthesis. We demonstrate BDQ hyper-susceptibility is common in INH-resistant clinical isolates. Collectively, these results highlight how altered bacterial physiology can impact drug efficacy in drug-resistant bacteria.

A historical perspective of biomedical explainable AI research

The black-box nature of most artificial intelligence (AI) models encourages the development of explainability methods to engender trust into the AI decision-making process. Such methods can be broadly categorized into two main types: post hoc explanations and inherently interpretable algorithms. We aimed at analyzing the possible associations between COVID-19 and the push of explainable AI (XAI) to the forefront of biomedical research. We automatically extracted from the PubMed database biomedical XAI studies related to concepts of causality or explainability and manually labeled 1,603 papers with respect to XAI categories. To compare the trends pre- and post-COVID-19, we fit a change point detection model and evaluated significant changes in publication rates. We show that the advent of COVID-19 in the beginning of 2020 could be the driving factor behind an increased focus concerning XAI, playing a crucial role in accelerating an already evolving trend. Finally, we present a discussion with future societal use and impact of XAI technologies and potential future directions for those who pursue fostering clinical trust with interpretable machine learning models.

Deep Neural Networks for Predicting Single-Cell Responses and Probability Landscapes

Engineering biology relies on the accurate prediction of cell responses. However, making these predictions is challenging for a variety of reasons, including the stochasticity of biochemical reactions, variability between cells, and incomplete information about underlying biological processes. Machine learning methods, which can model diverse input-output relationships without requiring a priori mechanistic knowledge, are an ideal tool for this task. For example, such approaches can be used to predict gene expression dynamics given time-series data of past expression history. To explore this application, we computationally simulated single-cell responses, incorporating different sources of noise and alternative genetic circuit designs. We showed that deep neural networks trained on these simulated data were able to correctly infer the underlying dynamics of a cell response even in the presence of measurement noise and stochasticity in the biochemical reactions. The training set size and the amount of past data provided as inputs both affected prediction quality, with cascaded genetic circuits that introduce delays requiring more past data. We also tested prediction performance on a bistable auto-activation circuit, finding that our initial method for predicting a single trajectory was fundamentally ill-suited for multimodal dynamics. To address this, we updated the network architecture to predict the entire distribution of future states, showing it could accurately predict bimodal expression distributions. Overall, these methods can be readily applied to the diverse prediction tasks necessary to predict and control a variety of biological circuits, a key aspect of many synthetic biology applications.

Analysis of Regulatory Mechanism of AcrB and CpxR on Colistin Susceptibility Based on Transcriptome and Metabolome of Salmonella Typhimurium

With the increasing and inappropriate use of colistin, the emerging colistin-resistant isolates have been frequently reported during the last few decades. Therefore, new potential targets and adjuvants to reverse colistin resistance are urgently needed. Our previous study has confirmed a marked increase of colistin susceptibility (16-fold compared to the wild-type Salmonella strain) of cpxR overexpression strain JSΔacrBΔcpxR::kan/pcpxR (simplified as JSΔΔ/pR). To searching for potential new drug targets, the transcriptome and metabolome analysis were carried out in this study. We found that the more susceptible strain JSΔΔ/pR displayed striking perturbations at both the transcriptomics and metabolomics levels. The virulence-related genes and colistin resistance-related genes (CRRGs) were significantly downregulated in JSΔΔ/pR. There were significant accumulation of citrate, α-ketoglutaric acid, and agmatine sulfate in JSΔΔ/pR, and exogenous supplement of them could synergistically enhance the bactericidal effect of colistin, indicating that these metabolites may serve as potential adjuvants for colistin therapy. Additionally, we also demonstrated that AcrB and CpxR could target the ATP and reactive oxygen species (ROS) generation, but not proton motive force (PMF) production pathway to potentiate antibacterial activity of colistin. Collectively, these findings have revealed several previously unknown mechanisms contributing to increased colistin susceptibility and identified potential targets and adjuvants for potentiating colistin treatment of Salmonella infections. IMPORTANCE Emergence of multidrug-resistant (MDR) Gram-negative (G-) bacteria have led to the reconsideration of colistin as the last-resort therapeutic option for health care-associated infections. Finding new drug targets and strategies against the spread of MDR G- bacteria are global challenges for the life sciences community and public health. In this paper, we demonstrated the more susceptibility strain JSΔΔ/pR displayed striking perturbations at both the transcriptomics and metabolomics levels and revealed several previously unknown regulatory mechanisms of AcrB and CpxR on the colistin susceptibility. Importantly, we found that exogenous supplement of citrate, α-ketoglutaric acid, and agmatine sulfate could synergistically enhance the bactericidal effect of colistin, indicating that these metabolites may serve as potential adjuvants for colistin therapy. These results provide a theoretical basis for finding potential new drug targets and adjuvants.

Profiling cell envelope-antibiotic interactions reveals vulnerabilities to β-lactams in a multidrug-resistant bacterium

The cell envelope of Gram-negative bacteria belonging to the Burkholderia cepacia complex (Bcc) presents unique restrictions to antibiotic penetration. As a consequence, Bcc species are notorious for causing recalcitrant multidrug-resistant infections in immunocompromised individuals. Here, we present the results of a genome-wide screen for cell envelope-associated resistance and susceptibility determinants in a Burkholderia cenocepacia clinical isolate. For this purpose, we construct a high-density, randomly-barcoded transposon mutant library and expose it to 19 cell envelope-targeting antibiotics. By quantifying relative mutant fitness with BarSeq, followed by validation with CRISPR-interference, we profile over a hundred functional associations and identify mediators of antibiotic susceptibility in the Bcc cell envelope. We reveal connections between β-lactam susceptibility, peptidoglycan synthesis, and blockages in undecaprenyl phosphate metabolism. The synergy of the β-lactam/β-lactamase inhibitor combination ceftazidime/avibactam is primarily mediated by inhibition of the PenB carbapenemase. In comparison with ceftazidime, avibactam more strongly potentiates the activity of aztreonam and meropenem in a panel of Bcc clinical isolates. Finally, we characterize in Bcc the iron and receptor-dependent activity of the siderophore-cephalosporin antibiotic, cefiderocol. Our work has implications for antibiotic target prioritization, and for using additional combinations of β-lactam/β-lactamase inhibitors that can extend the utility of current antibacterial therapies.

Integrating Omics Data in Genome-Scale Metabolic Modeling: A Methodological Perspective for Precision Medicine

Recent advancements in omics technologies have generated a wealth of biological data. Integrating these data within mathematical models is essential to fully leverage their potential. Genome-scale metabolic models (GEMs) provide a robust framework for studying complex biological systems. GEMs have significantly contributed to our understanding of human metabolism, including the intrinsic relationship between the gut microbiome and the host metabolism. In this review, we highlight the contributions of GEMs and discuss the critical challenges that must be overcome to ensure their reproducibility and enhance their prediction accuracy, particularly in the context of precision medicine. We also explore the role of machine learning in addressing these challenges within GEMs. The integration of omics data with GEMs has the potential to lead to new insights, and to advance our understanding of molecular mechanisms in human health and disease.

Leveraging artificial intelligence in the fight against infectious diseases

Despite advances in molecular biology, genetics, computation, and medicinal chemistry, infectious disease remains an ominous threat to public health. Addressing the challenges posed by pathogen outbreaks, pandemics, and antimicrobial resistance will require concerted interdisciplinary efforts. In conjunction with systems and synthetic biology, artificial intelligence (AI) is now leading to rapid progress, expanding anti-infective drug discovery, enhancing our understanding of infection biology, and accelerating the development of diagnostics. In this Review, we discuss approaches for detecting, treating, and understanding infectious diseases, underscoring the progress supported by AI in each case. We suggest future applications of AI and how it might be harnessed to help control infectious disease outbreaks and pandemics.

Biology-inspired graph neural network encodes reactome and reveals biochemical reactions of disease

Functional heterogeneity of healthy human tissues complicates interpretation of molecular studies, impeding precision therapeutic target identification and treatment. Considering this, we generated a graph neural network with Reactome-based architecture and trained it using 9,115 samples from Genotype-Tissue Expression (GTEx). Our graph neural network (GNN) achieves adjusted Rand index (ARI) = 0.7909, while a Resnet18 control model achieves ARI = 0.7781, on 370 held-out healthy human tissue samples from The Cancer Genome Atlas (TCGA), despite the Resnet18 using over 600 times the parameters. Our GNN also succeeds in separating 83 healthy skin samples from 95 lesional psoriasis samples, revealing that upregulation of 26S- and NUB1-mediated degradation of NEDD8, UBD, and their conjugates is central to the largest perturbed reaction network component in psoriasis. We show that our results are not discoverable using traditional differential expression and hypergeometric pathway enrichment analyses yet are supported by separate human multi-omics and small-molecule mouse studies, suggesting future molecular disease studies may benefit from similar GNN analytical approaches.

From genotype to phenotype: computational approaches for inferring microbial traits relevant to the food industry

When selecting microbial strains for the production of fermented foods, various microbial phenotypes need to be taken into account to achieve target product characteristics, such as biosafety, flavor, texture, and health-promoting effects. Through continuous advances in sequencing technologies, microbial whole-genome sequences of increasing quality can now be obtained both cheaper and faster, which increases the relevance of genome-based characterization of microbial phenotypes. Prediction of microbial phenotypes from genome sequences makes it possible to quickly screen large strain collections in silico to identify candidates with desirable traits. Several microbial phenotypes relevant to the production of fermented foods can be predicted using knowledge-based approaches, leveraging our existing understanding of the genetic and molecular mechanisms underlying those phenotypes. In the absence of this knowledge, data-driven approaches can be applied to estimate genotype-phenotype relationships based on large experimental datasets. Here, we review computational methods that implement knowledge- and data-driven approaches for phenotype prediction, as well as methods that combine elements from both approaches. Furthermore, we provide examples of how these methods have been applied in industrial biotechnology, with special focus on the fermented food industry.

Whole-cell modeling of E. coli colonies enables quantification of single-cell heterogeneity in antibiotic responses

Antibiotic resistance poses mounting risks to human health, as current antibiotics are losing efficacy against increasingly resistant pathogenic bacteria. Of particular concern is the emergence of multidrug-resistant strains, which has been rapid among Gram-negative bacteria such as Escherichia coli. A large body of work has established that antibiotic resistance mechanisms depend on phenotypic heterogeneity, which may be mediated by stochastic expression of antibiotic resistance genes. The link between such molecular-level expression and the population levels that result is complex and multi-scale. Therefore, to better understand antibiotic resistance, what is needed are new mechanistic models that reflect single-cell phenotypic dynamics together with population-level heterogeneity, as an integrated whole. In this work, we sought to bridge single-cell and population-scale modeling by building upon our previous experience in "whole-cell" modeling, an approach which integrates mathematical and mechanistic descriptions of biological processes to recapitulate the experimentally observed behaviors of entire cells. To extend whole-cell modeling to the "whole-colony" scale, we embedded multiple instances of a whole-cell E. coli model within a model of a dynamic spatial environment, allowing us to run large, parallelized simulations on the cloud that contained all the molecular detail of the previous whole-cell model and many interactive effects of a colony growing in a shared environment. The resulting simulations were used to explore the response of E. coli to two antibiotics with different mechanisms of action, tetracycline and ampicillin, enabling us to identify sub-generationally-expressed genes, such as the beta-lactamase ampC, which contributed greatly to dramatic cellular differences in steady-state periplasmic ampicillin and was a significant factor in determining cell survival.

Empowering drug off-target discovery with metabolic and structural analysis

Elucidating intracellular drug targets is a difficult problem. While machine learning analysis of omics data has been a promising approach, going from large-scale trends to specific targets remains a challenge. Here, we develop a hierarchic workflow to focus on specific targets based on analysis of metabolomics data and growth rescue experiments. We deploy this framework to understand the intracellular molecular interactions of the multi-valent dihydrofolate reductase-targeting antibiotic compound CD15-3. We analyse global metabolomics data utilizing machine learning, metabolic modelling, and protein structural similarity to prioritize candidate drug targets. Overexpression and in vitro activity assays confirm one of the predicted candidates, HPPK (folK), as a CD15-3 off-target. This study demonstrates how established machine learning methods can be combined with mechanistic analyses to improve the resolution of drug target finding workflows for discovering off-targets of a metabolic inhibitor.

Metagenomic next generation sequencing for studying antibiotic resistance genes in the environment

Bacterial antimicrobial resistance (AMR) is a persisting and growing threat to human health. Characterization of antibiotic resistance genes (ARGs) in the environment is important to understand and control ARG-associated microbial risks. Numerous challenges exist in monitoring ARGs in the environment, due to the extraordinary diversity of ARGs, low abundance of ARGs with respect to the complex environmental microbiomes, difficulties in linking ARGs with bacterial hosts by molecular methods, difficulties in achieving quantification and high throughput simultaneously, difficulties in assessing mobility potential of ARGs, and difficulties in determining the specific AMR determinant genes. Advances in the next generation sequencing (NGS) technologies and related computational and bioinformatic tools are facilitating rapid identification and characterization ARGs in genomes and metagenomes from environmental samples. This chapter discusses NGS-based strategies, including amplicon-based sequencing, whole genome sequencing, bacterial population-targeted metagenome sequencing, metagenomic NGS, quantitative metagenomic sequencing, and functional/phenotypic metagenomic sequencing. Current bioinformatic tools for analyzing sequencing data for studying environmental ARGs are also discussed.

Understanding Antimicrobial Resistance Using Genome-Scale Metabolic Modeling

The urgent necessity to fight antimicrobial resistance is universally recognized. In the search of new targets and strategies to face this global challenge, a promising approach resides in the study of the cellular response to antimicrobial exposure and on the impact of global cellular reprogramming on antimicrobial drugs' efficacy. The metabolic state of microbial cells has been shown to undergo several antimicrobial-induced modifications and, at the same time, to be a good predictor of the outcome of an antimicrobial treatment. Metabolism is a promising reservoir of potential drug targets/adjuvants that has not been fully exploited to date. One of the main problems in unraveling the metabolic response of cells to the environment resides in the complexity of such metabolic networks. To solve this problem, modeling approaches have been developed, and they are progressively gaining in popularity due to the huge availability of genomic information and the ease at which a genome sequence can be converted into models to run basic phenotype predictions. Here, we review the use of computational modeling to study the relationship between microbial metabolism and antimicrobials and the recent advances in the application of genome-scale metabolic modeling to the study of microbial responses to antimicrobial exposure.

Online Omics Platform Expedites Industrial Application of Halomonas bluephagenesis TD1.0

Multi-omic data mining has the potential to revolutionize synthetic biology especially in non-model organisms that have not been extensively studied. However, tangible engineering direction from computational analysis remains elusive due to the interpretability of large datasets and the difficulty in analysis for non-experts. New omics data are generated faster than our ability to use and analyse results effectively, resulting in strain development that proceeds through classic methods of trial-and-error without insight into complex cell dynamics. Here we introduce a user-friendly, interactive website hosting multi-omics data. Importantly, this new platform allows non-experts to explore questions in an industrially important chassis whose cellular dynamics are still largely unknown. The web platform contains a complete KEGG (Kyoto Encyclopedia of Genes and Genomes) pathway enrichment analysis derived from principal components analysis, an interactive bio-cluster heatmap analysis of genes, and the Halomonas TD1.0 genome-scale metabolic (GEM) model. As a case study of the effectiveness of this platform, we applied unsupervised machine learning to determine key differences between Halomonas bluephagenesis TD1.0 cultivated under varied conditions. Specifically, cell motility and flagella apparatus are identified to drive energy expenditure usage at different osmolarities, and predictions were verified experimentally using microscopy and fluorescence labelled flagella staining. As more omics projects are completed, this landing page will facilitate exploration and targeted engineering efforts of the robust, industrial chassis H bluephagenesis for researchers without extensive bioinformatics background.

Clinical Diagnostics of Bacterial Infections and Their Resistance to Antibiotics-Current State and Whole Genome Sequencing Implementation Perspectives

Antimicrobial resistance (AMR), defined as the ability of microorganisms to withstand antimicrobial treatment, is responsible for millions of deaths annually. The rapid spread of AMR across continents warrants systematic changes in healthcare routines and protocols. One of the fundamental issues with AMR spread is the lack of rapid diagnostic tools for pathogen identification and AMR detection. Resistance profile identification often depends on pathogen culturing and thus may last up to several days. This contributes to the misuse of antibiotics for viral infection, the use of inappropriate antibiotics, the overuse of broad-spectrum antibiotics, or delayed infection treatment. Current DNA sequencing technologies offer the potential to develop rapid infection and AMR diagnostic tools that can provide information in a few hours rather than days. However, these techniques commonly require advanced bioinformatics knowledge and, at present, are not suited for routine lab use. In this review, we give an overview of the AMR burden on healthcare, describe current pathogen identification and AMR screening methods, and provide perspectives on how DNA sequencing may be used for rapid diagnostics. Additionally, we discuss the common steps used for DNA data analysis, currently available pipelines, and tools for analysis. Direct, culture-independent sequencing has the potential to complement current culture-based methods in routine clinical settings. However, there is a need for a minimum set of standards in terms of evaluating the results generated. Additionally, we discuss the use of machine learning algorithms regarding pathogen phenotype detection (resistance/susceptibility to an antibiotic).

Antibiotics and Bacterial Resistance-A Short Story of an Endless Arms Race

Despite the undisputed development of medicine, antibiotics still serve as first-choice drugs for patients with infectious disorders. The widespread use of antibiotics results from a wide spectrum of their actions encompassing mechanisms responsible for: the inhibition of bacterial cell wall biosynthesis, the disruption of cell membrane integrity, the suppression of nucleic acids and/or proteins synthesis, as well as disturbances of metabolic processes. However, the widespread availability of antibiotics, accompanied by their overprescription, acts as a double-edged sword, since the overuse and/or misuse of antibiotics leads to a growing number of multidrug-resistant microbes. This, in turn, has recently emerged as a global public health challenge facing both clinicians and their patients. In addition to intrinsic resistance, bacteria can acquire resistance to particular antimicrobial agents through the transfer of genetic material conferring resistance. Amongst the most common bacterial resistance strategies are: drug target site changes, increased cell wall permeability to antibiotics, antibiotic inactivation, and efflux pumps. A better understanding of the interplay between the mechanisms of antibiotic actions and bacterial defense strategies against particular antimicrobial agents is crucial for developing new drugs or drug combinations. Herein, we provide a brief overview of the current nanomedicine-based strategies that aim to improve the efficacy of antibiotics.

MEMMAL: A tool for expanding large-scale mechanistic models with machine learned associations and big datasets

Computational models that can explain and predict complex sub-cellular, cellular, and tissue-level drug response mechanisms could speed drug discovery and prioritize patient-specific treatments (i.e., precision medicine). Some models are mechanistic with detailed equations describing known (or supposed) physicochemical processes, while some are statistical or machine learning-based approaches, that explain datasets but have no mechanistic or causal guarantees. These two types of modeling are rarely combined, missing the opportunity to explore possibly causal but data-driven new knowledge while explaining what is already known. Here, we explore combining machine learned associations with mechanistic models to develop computational models that could more fully represent cellular behavior. In this proposed MEMMAL (MEchanistic Modeling with MAchine Learning) framework, machine learning/statistical models built using omics datasets provide predictions for new interactions between genes and proteins where there is physicochemical uncertainty. These interactions are used as a basis for new reactions in mechanistic models. As a test case, we focused on incorporating novel IFNγ/PD-L1 related associations into a large-scale mechanistic model for cell proliferation and death to better recapitulate the recently released NIH LINCS Consortium MCF10A dataset and enable description of the cellular response to checkpoint inhibitor immunotherapies. This work is a template for combining big-data-inferred interactions with mechanistic models, which could be more broadly applicable for building multi-scale precision medicine and whole cell models.

Discovering Themes in Deep Brain Stimulation Research Using Explainable Artificial Intelligence

Deep brain stimulation is a treatment that controls symptoms by changing brain activity. The complexity of how to best treat brain dysfunction with deep brain stimulation has spawned research into artificial intelligence approaches. Machine learning is a subset of artificial intelligence that uses computers to learn patterns in data and has many healthcare applications, such as an aid in diagnosis, personalized medicine, and clinical decision support. Yet, how machine learning models make decisions is often opaque. The spirit of explainable artificial intelligence is to use machine learning models that produce interpretable solutions. Here, we use topic modeling to synthesize recent literature on explainable artificial intelligence approaches to extracting domain knowledge from machine learning models relevant to deep brain stimulation. The results show that patient classification (i.e., diagnostic models, precision medicine) is the most common problem in deep brain stimulation studies that employ explainable artificial intelligence. Other topics concern attempts to optimize stimulation strategies and the importance of explainable methods. Overall, this review supports the potential for artificial intelligence to revolutionize deep brain stimulation by personalizing stimulation protocols and adapting stimulation in real time.

A General Hybrid Modeling Framework for Systems Biology Applications: Combining Mechanistic Knowledge with Deep Neural Networks under the SBML Standard

In this paper, a computational framework is proposed that merges mechanistic modeling with deep neural networks obeying the Systems Biology Markup Language (SBML) standard. Over the last 20 years, the systems biology community has developed a large number of mechanistic models that are currently stored in public databases in SBML. With the proposed framework, existing SBML models may be redesigned into hybrid systems through the incorporation of deep neural networks into the model core, using a freely available python tool. The so-formed hybrid mechanistic/neural network models are trained with a deep learning algorithm based on the adaptive moment estimation method (ADAM), stochastic regularization and semidirect sensitivity equations. The trained hybrid models are encoded in SBML and uploaded in model databases, where they may be further analyzed as regular SBML models. This approach is illustrated with three well-known case studies: the Escherichia coli threonine synthesis model, the P58IPK signal transduction model, and the Yeast glycolytic oscillations model. The proposed framework is expected to greatly facilitate the widespread use of hybrid modeling techniques for systems biology applications.

Decoding the metabolic response of Escherichia coli for sensing trace heavy metals in water

Heavy metal contamination due to industrial and agricultural waste represents a growing threat to water supplies. Frequent and widespread monitoring for toxic metals in drinking and agricultural water sources is necessary to prevent their accumulation in humans, plants, and animals, which results in disease and environmental damage. Here, the metabolic stress response of bacteria is used to report the presence of heavy metal ions in water by transducing ions into chemical signals that can be fingerprinted using machine learning analysis of vibrational spectra. Surface-enhanced Raman scattering surfaces amplify chemical signals from bacterial lysate and rapidly generate large, reproducible datasets needed for machine learning algorithms to decode the complex spectral data. Classification and regression algorithms achieve limits of detection of 0.5 pM for As³⁺ and 6.8 pM for Cr⁶⁺, 100,000 times lower than the World Health Organization recommended limits, and accurately quantify concentrations of analytes across six orders of magnitude, enabling early warning of rising contaminant levels. Trained algorithms are generalizable across water samples with different impurities; water quality of tap water and wastewater was evaluated with 92% accuracy.

Mechanism of Action of Isopropoxy Benzene Guanidine against Multidrug-Resistant Pathogens

The increasing emergence of antibiotic resistance is an urgent threat to global health care; thus, there is a need for new therapeutics. Guanidine is the preferred functional group for antimicrobial design and development. Herein, the potential antibacterial activity of the guanidine derivative isopropoxy benzene guanidine (IBG) against multidrug-resistant (MDR) bacteria was discovered. The synergistic antibacterial activity of IBG and colistin was determined by checkerboard assay, time-killing curve, and mouse experiments. The antibacterial mechanism of IBG was verified in fluorescent probe experiments, intracellular oxidative phosphorylation assays, and transcriptome analysis. The results showed that IBG displays efficient antibacterial activity against Gram-positive pathogens and Gram-negative pathogens with permeabilized outer membranes. Further mechanistic studies showed that IBG triggers cytoplasmic membrane damage by binding to phosphatidylglycerol and cardiolipin, leading to the dissipation of proton motive force and accumulation of intracellular ATP. IBG combined with low levels of colistin enhances bacterial outer membrane permeability and increases the accumulation of reactive oxygen species, as further evidenced by transcriptome analysis. Furthermore, the efficacy of IBG with colistin against MDR Escherichia coli in three infection models was demonstrated. Together, these results suggest that IBG is a promising adjuvant of colistin, providing an alternative approach to address the prevalent infections caused by MDR Gram-negative pathogens. IMPORTANCE As antibiotic discovery stagnates, the world is facing a growing menace from the emergence of bacteria that are resistant to almost all available antibiotics. The key to winning this race is to explore distinctive mechanisms of antibiotics. Thus, novel efficient antibacterial agents and alternative strategies are urgently required to fill the void in antibiotic development. Compared with the large amount of money and time required to develop new agents, the antibiotic adjuvant strategy is a promising approach to inhibit bacterial resistance and increase killing of bacteria. In this study, we found that the guanidine derivatives IBG not only displayed efficient antibacterial activities against Gram-positive bacteria but also restored colistin susceptibility of Gram-negative pathogens as an antibiotic adjuvant. More in-depth study showed that IBG is a potential lead to overcome antibiotic resistance, providing new insight into future antibiotic discovery and development.

Genome-scale model of Pseudomonas aeruginosa metabolism unveils virulence and drug potentiation

Pseudomonas aeruginosa is one of the leading causes of hospital-acquired infections. To decipher the metabolic mechanisms associated with virulence and antibiotic resistance, we have developed an updated genome-scale model (GEM) of P. aeruginosa. The model (iSD1509) is an extensively curated, three-compartment, and mass-and-charge balanced BiGG model containing 1509 genes, the largest gene content for any P. aeruginosa GEM to date. It is the most accurate with prediction accuracies as high as 92.4% (gene essentiality) and 93.5% (substrate utilization). In iSD1509, we newly added a recently discovered pathway for ubiquinone-9 biosynthesis which is required for anaerobic growth. We used a modified iSD1509 to demonstrate the role of virulence factor (phenazines) in the pathogen survival within biofilm/oxygen-limited condition. Further, the model can mechanistically explain the overproduction of a drug susceptibility biomarker in the P. aeruginosa mutants. Finally, we use iSD1509 to demonstrate the drug potentiation by metabolite supplementation, and elucidate the mechanisms behind the phenotype, which agree with experimental results.

Comparative Metabolome Analyses of Ivermectin-Resistant and -Susceptible Strains of Haemonchus contortus

Resistance to anthelmintics such as ivermectin (IVM) is currently a major problem in the treatment of Haemonchus contortus, an important parasitic nematode of small ruminants. Although many advances have been made in understanding the IVM resistance mechanism, its exact mechanism remains unclear for H. contortus. Therefore, understanding the resistance mechanism becomes increasingly important for controlling haemonchosis. Recent research showed that the metabolic state of bacteria influences their susceptibility to antibiotics. However, little information is available on the roles of metabolites and metabolic pathways in IVM resistance of H. contortus. In this study, comparative analyses of the metabolomics of IVM-susceptible and -resistant adult H. contortus worms were carried out to explore the role of H. contortus metabolism in IVM resistance. In total, 705 metabolites belonging to 42 categories were detected, and 86 differential metabolites (17 upregulated and 69 downregulated) were identified in the IVM-resistant strain compared to the susceptible one. A KEGG pathway analysis showed that these 86 differential metabolites were enriched in 42 pathways that mainly included purine metabolism; the biosynthesis of amino acids; glycine, serine, and threonine metabolism; and cysteine and methionine metabolism. These results showed that amino acid metabolism may be mediated by the uptake of IVM and related with IVM resistance in H. contortus. This study contributes to our understanding of the mechanisms of IVM resistance and may provide effective approaches to manage infection by resistant strains of H. contortus.

Uracil restores susceptibility of methicillin-resistant Staphylococcus aureus to aminoglycosides through metabolic reprogramming

Background: Methicillin-resistant Staphylococcus aureus (MRSA) has now become a major nosocomial pathogen bacteria and resistant to many antibiotics. Therefore, Development of novel approaches to combat the disease is especially important. The present study aimed to provide a novel approach involving the use of nucleotide-mediated metabolic reprogramming to tackle intractable methicillin-resistant S. aureus (MRSA) infections. Objective: This study aims to explore the bacterial effects and mechanism of uracil and gentamicin in S. aureus. Methods: Antibiotic bactericidal assays was used to determine the synergistic bactericidal effect of uracil and gentamicin. How did uracil regulate bacterial metabolism including the tricarboxylic acid (TCA) cycle by GC-MS-based metabolomics. Next, genes and activity of key enzymes in the TCA cycle, PMF, and intracellular aminoglycosides were measured. Finally, bacterial respiration, reactive oxygen species (ROS), and ATP levels were also assayed in this study. Results: In the present study, we found that uracil could synergize with aminoglycosides to kill MRSA (USA300) by 400-fold. Reprogramming metabolomics displayed uracil reprogrammed bacterial metabolism, especially enhanced the TCA cycle to elevate NADH production and proton motive force, thereby promoting the uptake of antibiotics. Furthermore, uracil increased cellular respiration and ATP production, resulting the generation of ROS. Thus, the combined activity of uracil and antibiotics induced bacterial death. Inhibition of the TCA cycle or ROS production could attenuate bactericidal efficiency. Moreover, uracil exhibited bactericidal activity in cooperation with aminoglycosides against other pathogenic bacteria. In a mouse mode of MRSA infection, the combination of gentamicin and uracil increased the survival rate of infected mice. Conclusion: Our results suggest that uracil enhances the activity of bactericidal antibiotics to kill Gram-positive bacteria by modulating bacterial metabolism.

SBML2HYB: a Python interface for SBML compatible hybrid modeling

SUMMARY

Here, we present sbml2hyb, an easy-to-use standalone Python tool that facilitates the conversion of existing mechanistic models of biological systems in Systems Biology Markup Language (SBML) into hybrid semiparametric models that combine mechanistic functions with machine learning (ML). The so-formed hybrid models can be trained and stored back in databases in SBML format. The tool supports a user-friendly export interface with an internal format validator. Two case studies illustrate the use of the sbml2hyb tool. Additionally, we describe HMOD, a new model format designed to support and facilitate hybrid models building. It aggregates the mechanistic model information with the ML information and follows as close as possible the SBML rules. We expect the sbml2hyb tool and HMOD to greatly facilitate the widespread usage of hybrid modeling techniques for biological systems analysis.

AVAILABILITY AND IMPLEMENTATION

The Python interface, source code and the example models used for the case studies are accessible at: https://github.com/r-costa/sbml2hyb.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.