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

Artificial intelligence applications in predicting the behavior of gastrointestinal cancers in pathology

Recent research has provided a wealth of data supporting the application of artificial intelligence (AI)-based applications in routine pathology practice. Indeed, it is clear that these methods can significantly support an accurate and rapid diagnosis by eliminating errors, increasing reliability, and improving workflow. In addition, the effectiveness of AI in the pathological evaluation of prognostic parameters associated with behavior, course, and treatment in many types of tumors has also been noted. Regarding gastrointestinal system (GIS) cancers, the contribution of AI methods to pathological diagnosis has been investigated in many studies. On the other hand, studies focusing on AI applications in evaluating parameters to determine tumor behavior are relatively few. For this purpose, the potential of AI models has been studied over a broad spectrum, from tumor subtyping to the identification of new digital biomarkers. The capacity of AI to infer genetic alterations of cancer tissues from digital slides has been demonstrated. Although current data suggest the merit of AI-based approaches in assessing tumor behavior in GIS cancers, a wide range of challenges still need to be solved, from laboratory infrastructure to improving the robustness of algorithms, before incorporating AI applications into real-life GIS pathology practice. This review aims to present data from AI applications in evaluating pathological parameters related to the behavior of GIS cancer with an overview of the opportunities and challenges encountered in implementing AI in pathology.

RNA-Seq-based transcriptome analysis of methicillin-resistant Staphylococcus aureus growth inhibition by propionate

Staphylococcus aureus is a pathogen that causes a variety of infectious diseases such as pneumonia, endocarditis, and septic shock. Methicillin-resistant S. aureus (MRSA) evades virtually all available treatments, creating the need for an alternative control strategy. Although we previously demonstrated the inhibitory effect of sodium propionate (NaP) on MRSA, the regulatory mechanism of this effect remains unclear. In this study, we investigated the regulatory mechanism responsible for the inhibitory effect of NaP on MRSA using RNA-Seq analysis. Total RNAs were isolated from non-treated and 50 mM NaP-treated S. aureus USA300 for 3 h and transcriptional profiling was conducted by RNA-Seq analysis. A total of 171 differentially expressed genes (DEGs) with log₂ fold change ≥2 and p < 0.05 was identified in the NaP treatment group compared with the control group. Among the 171 genes, 131 were up-regulated and 40 were down-regulated. Upon gene ontology (GO) annotation analysis, total 26 specific GO terms in "Biological process," "Molecular function," and "Cellular component" were identified in MRSA treated with NaP for 3 h. "Purine metabolism"; "riboflavin metabolism"; and "glycine, serine, and threonine metabolism" were identified as major altered metabolic pathways among the eight significantly enriched KEGG pathways in MRSA treated with NaP. Furthermore, the MRSA strains deficient in purF, ilvA, ribE, or ribA, which were the up-regulated DEGs in the metabolic pathways, were more susceptible to NaP than wild-type MRSA. Collectively, these results demonstrate that NaP attenuates MRSA growth by altering its metabolic pathways, suggesting that NaP can be used as a potential bacteriostatic agent for prevention of MRSA infection.

Coordination of CcpA and CodY Regulators in Staphylococcus aureus USA300 Strains

The complex cross talk between metabolism and gene regulatory networks makes it difficult to untangle individual constituents and study their precise roles and interactions. To address this issue, we modularized the transcriptional regulatory network (TRN) of the Staphylococcus aureus USA300 strain by applying independent component analysis (ICA) to 385 RNA sequencing samples. We then combined the modular TRN model with a metabolic model to study the regulation of carbon and amino acid metabolism. Our analysis showed that regulation of central carbon metabolism by CcpA and amino acid biosynthesis by CodY are closely coordinated. In general, S. aureus increases the expression of CodY-regulated genes in the presence of preferred carbon sources such as glucose. This transcriptional coordination was corroborated by metabolic model simulations that also showed increased amino acid biosynthesis in the presence of glucose. Further, we found that CodY and CcpA cooperatively regulate the expression of ribosome hibernation-promoting factor, thus linking metabolic cues with translation. In line with this hypothesis, expression of CodY-regulated genes is tightly correlated with expression of genes encoding ribosomal proteins. Together, we propose a coarse-grained model where expression of S. aureus genes encoding enzymes that control carbon flux and nitrogen flux through the system is coregulated with expression of translation machinery to modularly control protein synthesis. While this work focuses on three key regulators, the full TRN model we present contains 76 total independently modulated sets of genes, each with the potential to uncover other complex regulatory structures and interactions. IMPORTANCE Staphylococcus aureus is a versatile pathogen with an expanding antibiotic resistance profile. The biology underlying its clinical success emerges from an interplay of many systems such as metabolism and gene regulatory networks. This work brings together models for these two systems to establish fundamental principles governing the regulation of S. aureus central metabolism and protein synthesis. Studies of these fundamental biological principles are often confined to model organisms such as Escherichia coli. However, expanding these models to pathogens can provide a framework from which complex and clinically important phenotypes such as virulence and antibiotic resistance can be better understood. Additionally, the expanded gene regulatory network model presented here can deconvolute the biology underlying other important phenotypes in this pathogen.

Fighting antibiotic resistance-strategies and (pre)clinical developments to find new antibacterials

Antibacterial resistance is one of the greatest threats to human health. The development of new therapeutics against bacterial pathogens has slowed drastically since the approvals of the first antibiotics in the early and mid-20^th century. Most of the currently investigated drug leads are modifications of approved antibacterials, many of which are derived from natural products. In this review, we highlight the challenges, advancements and current standing of the clinical and preclinical antibacterial research pipeline. Additionally, we present novel strategies for rejuvenating the discovery process and advocate for renewed and enthusiastic investment in the antibacterial discovery pipeline.

Growth productivity as a determinant of the inoculum effect for bactericidal antibiotics

Understanding the mechanisms by which populations of bacteria resist antibiotics has implications in evolution, microbial ecology, and public health. The inoculum effect (IE), where antibiotic efficacy declines as the density of a bacterial population increases, has been observed for multiple bacterial species and antibiotics. Several mechanisms to account for IE have been proposed, but most lack experimental evidence or cannot explain IE for multiple antibiotics. We show that growth productivity, the combined effect of growth and metabolism, can account for IE for multiple bactericidal antibiotics and bacterial species. Guided by flux balance analysis and whole-genome modeling, we show that the carbon source supplied in the growth medium determines growth productivity. If growth productivity is sufficiently high, IE is eliminated. Our results may lead to approaches to reduce IE in the clinic, help standardize the analysis of antibiotics, and further our understanding of how bacteria evolve resistance.

Deep Transfer Learning Enables Robust Prediction of Antimicrobial Resistance for Novel Antibiotics

Antimicrobial resistance (AMR) has become one of the serious global health problems, threatening the effective treatment of a growing number of infections. Machine learning and deep learning show great potential in rapid and accurate AMR predictions. However, a large number of samples for the training of these models is essential. In particular, for novel antibiotics, limited training samples and data imbalance hinder the models’ generalization performance and overall accuracy. We propose a deep transfer learning model that can improve model performance for AMR prediction on small, imbalanced datasets. As our approach relies on transfer learning and secondary mutations, it is also applicable to novel antibiotics and emerging resistances in the future and enables quick diagnostics and personalized treatments.

Interpretable machine learning methods for predictions in systems biology from omics data

Machine learning has become a powerful tool for systems biologists, from diagnosing cancer to optimizing kinetic models and predicting the state, growth dynamics, or type of a cell. Potential predictions from complex biological data sets obtained by “omics” experiments seem endless, but are often not the main objective of biological research. Often we want to understand the molecular mechanisms of a disease to develop new therapies, or we need to justify a crucial decision that is derived from a prediction. In order to gain such knowledge from data, machine learning models need to be extended. A recent trend to achieve this is to design “interpretable” models. However, the notions around interpretability are sometimes ambiguous, and a universal recipe for building well-interpretable models is missing. With this work, we want to familiarize systems biologists with the concept of model interpretability in machine learning. We consider data sets, data preparation, machine learning methods, and software tools relevant to omics research in systems biology. Finally, we try to answer the question: “What is interpretability?” We introduce views from the interpretable machine learning community and propose a scheme for categorizing studies on omics data. We then apply these tools to review and categorize recent studies where predictive machine learning models have been constructed from non-sequential omics data.

Metabolomics and modelling approaches for systems metabolic engineering

Metabolic engineering involves the manipulation of microbes to produce desirable compounds through genetic engineering or synthetic biology approaches. Metabolomics involves the quantitation of intracellular and extracellular metabolites, where mass spectrometry and nuclear magnetic resonance based analytical instrumentation are often used. Here, the experimental designs, sample preparations, metabolite quenching and extraction are essential to the quantitative metabolomics workflow. The resultant metabolomics data can then be used with computational modelling approaches, such as kinetic and constraint-based modelling, to better understand underlying mechanisms and bottlenecks in the synthesis of desired compounds, thereby accelerating research through systems metabolic engineering. Constraint-based models, such as genome scale models, have been used successfully to enhance the yield of desired compounds from engineered microbes, however, unlike kinetic or dynamic models, constraint-based models do not incorporate regulatory effects. Nevertheless, the lack of time-series metabolomic data generation has hindered the usefulness of dynamic models till today. In this review, we show that improvements in automation, dynamic real-time analysis and high throughput workflows can drive the generation of more quality data for dynamic models through time-series metabolomics data generation. Spatial metabolomics also has the potential to be used as a complementary approach to conventional metabolomics, as it provides information on the localization of metabolites. However, more effort must be undertaken to identify metabolites from spatial metabolomics data derived through imaging mass spectrometry, where machine learning approaches could prove useful. On the other hand, single-cell metabolomics has also seen rapid growth, where understanding cell-cell heterogeneity can provide more insights into efficient metabolic engineering of microbes. Moving forward, with potential improvements in automation, dynamic real-time analysis, high throughput workflows, and spatial metabolomics, more data can be produced and studied using machine learning algorithms, in conjunction with dynamic models, to generate qualitative and quantitative predictions to advance metabolic engineering efforts.

Intracellular acidification is a hallmark of thymineless death in E. coli

Thymidine starvation causes rapid cell death. This enigmatic process known as thymineless death (TLD) is the underlying killing mechanism of diverse antimicrobial and antineoplastic drugs. Despite decades of investigation, we still lack a mechanistic understanding of the causal sequence of events that culminate in TLD. Here, we used a diverse set of unbiased approaches to systematically determine the genetic and regulatory underpinnings of TLD in Escherichia coli. In addition to discovering novel genes in previously implicated pathways, our studies revealed a critical and previously unknown role for intracellular acidification in TLD. We observed that a decrease in cytoplasmic pH is a robust early event in TLD across different genetic backgrounds. Furthermore, we show that acidification is a causal event in the death process, as chemical and genetic perturbations that increase intracellular pH substantially reduce killing. We also observe a decrease in intracellular pH in response to exposure to the antibiotic gentamicin, suggesting that intracellular acidification may be a common mechanistic step in the bactericidal effects of other antibiotics.

Reactive metabolic byproducts contribute to antibiotic lethality under anaerobic conditions

Understanding how bactericidal antibiotics kill bacteria remains an open question. Previous work has proposed that primary drug-target corruption leads to increased energetic demands, resulting in the generation of reactive metabolic byproducts (RMBs), particularly reactive oxygen species, that contribute to antibiotic-induced cell death. Studies have challenged this hypothesis by pointing to antibiotic lethality under anaerobic conditions. Here, we show that treatment of Escherichia coli with bactericidal antibiotics under anaerobic conditions leads to changes in the intracellular concentrations of central carbon metabolites, as well as the production of RMBs, particularly reactive electrophilic species (RES). We show that antibiotic treatment results in DNA double-strand breaks and membrane damage and demonstrate that antibiotic lethality under anaerobic conditions can be decreased by RMB scavengers, which reduce RES accumulation and mitigate associated macromolecular damage. This work indicates that RMBs, generated in response to antibiotic-induced energetic demands, contribute in part to antibiotic lethality under anaerobic conditions.

Machine learning in the identification, prediction and exploration of environmental toxicology: Challenges and perspectives

Over the past few decades, data-driven machine learning (ML) has distinguished itself from hypothesis-driven studies and has recently received much attention in environmental toxicology. However, the use of ML in environmental toxicology remains in the early stages, with knowledge gaps, technical bottlenecks in data quality, high-dimensional/heterogeneous/small-sample data analysis and model interpretability, and a lack of an in-depth understanding of environmental toxicology. Given the above problems, we review the recent progress in the literature and highlight state-of-the-art toxicological studies using ML (such as learning and predicting toxicity in complicated biosystems and multiple-factor environmental scenarios of long-term and large-scale pollution). Beyond predicting simple biological endpoints by integrating untargeted omics and adverse outcome pathways, ML development should focus on revealing toxicological mechanisms. The integration of data-driven ML with other methods (e.g., omics analysis and adverse outcome pathway frameworks) endows ML with widely promising application in revealing toxicological mechanisms. High-quality databases and interpretable algorithms are urgently needed for toxicology and environmental science. Addressing the core issues and future challenges for ML in this review may narrow the knowledge gap between environmental toxicity and computational science and facilitate the control of environmental risk in the future.

A Deep Neural Network for Gastric Cancer Prognosis Prediction Based on Biological Information Pathways

Background

Gastric cancer (GC) is one of the deadliest cancers in the world, with a 5-year overall survival rate of lower than 20% for patients with advanced GC. Genomic information is now frequently employed for precision cancer treatment due to the rapid advancements of high-throughput sequencing technologies. As a result, integrating multiomics data to construct predictive models for the GC patient prognosis is critical for tailored medical care.

Results

In this study, we integrated multiomics data to design a biological pathway-based gastric cancer sparse deep neural network (GCS-Net) by modifying the P-NET model for long-term survival prediction of GC. The GCS-Net showed higher accuracy (accuracy = 0.844), area under the curve (AUC = 0.807), and F1 score (F1 = 0.913) than traditional machine learning models. Furthermore, the GCS-Net not only enables accurate patient survival prognosis but also provides model interpretability capabilities lacking in most traditional deep neural networks to describe the complex biological process of prognosis. The GCS-Net suggested the importance of genes (UBE2C, JAK2, RAD21, CEP250, NUP210, PTPN1, CDC27, NINL, NUP188, and PLK4) and biological pathways (Mitotic Anaphase, Resolution of Sister Chromatid Cohesion, and SUMO E3 ligases) to GC, which is consistent with the results revealed in biological- and medical-related studies of GC.

Conclusion

The GCS-Net is an interpretable deep neural network built using biological pathway information whose structure represents a nonlinear hierarchical representation of genes and biological pathways. It can not only accurately predict the prognosis of GC patients but also suggest the importance of genes and biological pathways. The GCS-Net opens up new avenues for biological research and could be adapted for other cancer prediction and discovery activities as well.

The antimicrobial peptide LI14 combats multidrug-resistant bacterial infections

The prevalence of multidrug-resistant (MDR) pathogens raises public fears of untreatable infections and represents a huge health risk. There is an urgent need to exploit novel antimicrobial agents. Due to the unique mechanisms, antimicrobial peptides (AMPs) with a low probability to achieve resistance are regarded as potential antibiotic alternatives to address this issue. Herein, we develop a panel of synthetic peptide compounds with novel structures based on the database filters technology (DFT), and the lead peptide LI14 shows potent antibacterial activity against all tested drug-resistant bacteria. LI14 exhibits rapid bactericidal activity and excellent anti-biofilm and -persisters activity, simultaneously showing a low propensity to induce resistance. Moreover, LI14 shows tolerance against pH, temperatures, and pepsin treatment, and no detectable toxicity both in vitro and in vivo. Mechanistic studies revealed that LI14 induces membrane damage by targeting bacterial-specific membrane components and dissipates the proton motive force (PMF), thereby resulting in metabolic perturbations and the accumulation of toxic metabolic products. Furthermore, LI14 sensitizes clinically relevant antibiotics against MDR bacteria. In animal models of infection, LI14 or combined with antibiotics are effective against drug-resistant pathogens. These findings suggest that LI14 is a promising antibiotic candidate to tackle MDR bacterial infections.

A synthetic peptide LI14 demonstrates potent antibacterial activity against drug-resistant bacteria in vitro and in vivo by inducing membrane damage and disrupting membrane potential leading to metabolic perturbation.

Systematic profiling of the chicken gut microbiome reveals dietary supplementation with antibiotics alters expression of multiple microbial pathways with minimal impact on community structure

BACKGROUND

The emergence of antimicrobial resistance is a major threat to global health and has placed pressure on the livestock industry to eliminate the use of antibiotic growth promotants (AGPs) as feed additives. To mitigate their removal, efficacious alternatives are required. AGPs are thought to operate through modulating the gut microbiome to limit opportunities for colonization by pathogens, increase nutrient utilization, and reduce inflammation. However, little is known concerning the underlying mechanisms. Previous studies investigating the effects of AGPs on the poultry gut microbiome have largely focused on 16S rDNA surveys based on a single gastrointestinal (GI) site, diet, and/or timepoint, resulting in an inconsistent view of their impact on community composition.

METHODS

In this study, we perform a systematic investigation of both the composition and function of the chicken gut microbiome, in response to AGPs. Birds were raised under two different diets and AGP treatments, and 16S rDNA surveys applied to six GI sites sampled at three key timepoints of the poultry life cycle. Functional investigations were performed through metatranscriptomics analyses and metabolomics.

RESULTS

Our study reveals a more nuanced view of the impact of AGPs, dependent on age of bird, diet, and intestinal site sampled. Although AGPs have a limited impact on taxonomic abundances, they do appear to redefine influential taxa that may promote the exclusion of other taxa. Microbiome expression profiles further reveal a complex landscape in both the expression and taxonomic representation of multiple pathways including cell wall biogenesis, antimicrobial resistance, and several involved in energy, amino acid, and nucleotide metabolism. Many AGP-induced changes in metabolic enzyme expression likely serve to redirect metabolic flux with the potential to regulate bacterial growth or produce metabolites that impact the host.

CONCLUSIONS

As alternative feed additives are developed to mimic the action of AGPs, our study highlights the need to ensure such alternatives result in functional changes that are consistent with site-, age-, and diet-associated taxa. The genes and pathways identified in this study are therefore expected to drive future studies, applying tools such as community-based metabolic modeling, focusing on the mechanistic impact of different dietary regimes on the microbiome. Consequently, the data generated in this study will be crucial for the development of next-generation feed additives targeting gut health and poultry production. Video Abstract.

Antimicrobial Resistance and Its Spread Is a Global Threat

Antimicrobial resistance (AMR) is a challenge to human wellbeing the world over and is one of the more serious public health concerns. AMR has the potential to emerge as a serious healthcare threat if left unchecked, and could put into motion another pandemic. This establishes the need for the establishment of global health solutions around AMR, taking into account microdata from different parts of the world. The positive influences in this regard could be establishing conducive social norms, charting individual and group behavior practices that favor global human health, and lastly, increasing collective awareness around the need for such action. Apart from being an emerging threat in the clinical space, AMR also increases treatment complexity, posing a real challenge to the existing guidelines around the management of antibiotic resistance. The attribute of resistance development has been linked to many genetic elements, some of which have complex transmission pathways between microbes. Beyond this, new mechanisms underlying the development of AMR are being discovered, making this field an important aspect of medical microbiology. Apart from the genetic aspects of AMR, other practices, including misdiagnosis, exposure to broad-spectrum antibiotics, and lack of rapid diagnosis, add to the creation of resistance. However, upgrades and innovations in DNA sequencing technologies with bioinformatics have revolutionized the diagnostic industry, aiding the real-time detection of causes of AMR and its elements, which are important to delineating control and prevention approaches to fight the threat.

In silico genome-scale metabolic modeling and in vitro static time-kill studies of exogenous metabolites alone and with polymyxin B against Klebsiella pneumoniae

Multidrug-resistant (MDR) Klebsiella pneumoniae is a top-prioritized Gram-negative pathogen with a high incidence in hospital-acquired infections. Polymyxins have resurged as a last-line therapy to combat Gram-negative “superbugs”, including MDR K. pneumoniae. However, the emergence of polymyxin resistance has increasingly been reported over the past decades when used as monotherapy, and thus combination therapy with non-antibiotics (e.g., metabolites) becomes a promising approach owing to the lower risk of resistance development. Genome-scale metabolic models (GSMMs) were constructed to delineate the altered metabolism of New Delhi metallo-β-lactamase- or extended spectrum β-lactamase-producing K. pneumoniae strains upon addition of exogenous metabolites in media. The metabolites that caused significant metabolic perturbations were then selected to examine their adjuvant effects using in vitro static time–kill studies. Metabolic network simulation shows that feeding of 3-phosphoglycerate and ribose 5-phosphate would lead to enhanced central carbon metabolism, ATP demand, and energy consumption, which is converged with metabolic disruptions by polymyxin treatment. Further static time–kill studies demonstrated enhanced antimicrobial killing of 10 mM 3-phosphoglycerate (1.26 and 1.82 log₁₀ CFU/ml) and 10 mM ribose 5-phosphate (0.53 and 0.91 log₁₀ CFU/ml) combination with 2 mg/L polymyxin B against K. pneumoniae strains. Overall, exogenous metabolite feeding could possibly improve polymyxin B activity via metabolic modulation and hence offers an attractive approach to enhance polymyxin B efficacy. With the application of GSMM in bridging the metabolic analysis and time–kill assay, biological insights into metabolite feeding can be inferred from comparative analyses of both results. Taken together, a systematic framework has been developed to facilitate the clinical translation of antibiotic-resistant infection management.

Genome-scale metabolic network models: from first-generation to next-generation

Over the last two decades, thousands of genome-scale metabolic network models (GSMMs) have been constructed. These GSMMs have been widely applied in various fields, ranging from network interaction analysis, to cell phenotype prediction. However, due to the lack of constraints, the prediction accuracy of first-generation GSMMs was limited. To overcome these limitations, the next-generation GSMMs were developed by integrating omics data, adding constrain condition, integrating different biological models, and constructing whole-cell models. Here, we review recent advances of GSMMs from the first generation to the next generation. Then, we discuss the major application of GSMMs in industrial biotechnology, such as predicting phenotypes and guiding metabolic engineering. In addition, human health applications, including understanding biological mechanisms, discovering biomarkers and drug targets, are also summarized. Finally, we address the challenges and propose new trend of GSMMs. KEY POINTS: •This mini-review updates the literature on almost all published GSMMs since 1999. •Detailed insights into the development of the first- and next-generation GSMMs. •The application of GSMMs is summarized, and the prospects of integrating machine learning are emphasized.

Deficiency of GntR Family Regulator MSMEG_5174 Promotes Mycobacterium smegmatis Resistance to Aminoglycosides via Manipulating Purine Metabolism

The increasing incidence of drug-resistant tuberculosis is still an emergency for global public health and a major obstacle to tuberculosis treatment. Therefore, deciphering the novel mechanisms of mycobacterial antibiotic resistance is crucial for combatting the rapid emergence of drug-resistant strains. In this study, we identified an unexpected role of Mycobacterium smegmatis GntR family transcriptional regulator MSMEG_5174 and its homologous gene Mycobacterium tuberculosis Rv1152 in aminoglycoside antibiotic resistance. Deficiency of MSMEG_5174 rendered Mycobacterium smegmatis highly resistant to aminoglycoside antibiotic treatment, and ectopic expression of Rv1152 in MSMEG_5174 mutants restored antibiotic-induced bacterial killing. We further demonstrated that MSMEG_5174 negatively regulates the expression of purine metabolism-related genes and the accumulation of purine metabolites. Moreover, overexpression of xanthine dehydrogenase MSMEG_0871 or xanthine treatment elicited a significant decrease in aminoglycoside antibiotic lethality for Mycobacterium smegmatis. Together, our findings revealed MSMEG_5174 as a metabolic regulator and hint toward unexplored crosstalk between purine metabolism and antibiotic resistance.

Exogenous metabolite feeding on altering antibiotic susceptibility in Gram-negative bacteria through metabolic modulation: a review

BACKGROUND

The rise of antimicrobial resistance at an alarming rate is outpacing the development of new antibiotics. The worrisome trends of multidrug-resistant Gram-negative bacteria have enormously diminished existing antibiotic activity. Antibiotic treatments may inhibit bacterial growth or lead to induce bacterial cell death through disruption of bacterial metabolism directly or indirectly. In light of this, it is imperative to have a thorough understanding of the relationship of bacterial metabolism with antimicrobial activity and leverage the underlying principle towards development of novel and effective antimicrobial therapies.

OBJECTIVE

Herein, we explore studies on metabolic analyses of Gram-negative pathogens upon antibiotic treatment. Metabolomic studies revealed that antibiotic therapy caused changes of metabolites abundance and perturbed the bacterial metabolism. Following this line of thought, addition of exogenous metabolite has been employed in in vitro, in vivo and in silico studies to activate the bacterial metabolism and thus potentiate the antibiotic activity.

KEY SCIENTIFIC CONCEPTS OF REVIEW

Exogenous metabolites were discovered to cause metabolic modulation through activation of central carbon metabolism and cellular respiration, stimulation of proton motive force, increase of membrane potential, improvement of host immune protection, alteration of gut microbiome, and eventually facilitating antibiotic killing. The use of metabolites as antimicrobial adjuvants may be a promising approach in the fight against multidrug-resistant pathogens.

A flux-based machine learning model to simulate the impact of pathogen metabolic heterogeneity on drug interactions

Drug combinations are a promising strategy to counter antibiotic resistance. However, current experimental and computational approaches do not account for the entire complexity involved in combination therapy design, such as the effect of pathogen metabolic heterogeneity, changes in the growth environment, drug treatment order, and time interval. To address these limitations, we present a comprehensive approach that uses genome-scale metabolic modeling and machine learning to guide combination therapy design. Our mechanistic approach (a) accommodates diverse data types, (b) accounts for time- and order-specific interactions, and (c) accurately predicts drug interactions in various growth conditions and their robustness to pathogen metabolic heterogeneity. Our approach achieved high accuracy (area under the receiver operating curve (AUROC) = 0.83 for synergy, AUROC = 0.98 for antagonism) in predicting drug interactions for Escherichia coli cultured in 57 metabolic conditions based on experimental validation. The entropy in bacterial metabolic response was predictive of combination therapy outcomes across time scales and growth conditions. Simulation of metabolic heterogeneity using population FBA identified two subpopulations of E. coli cells defined by the levels of three proteins (eno, fadB, and fabD) in glycolysis and lipid metabolism that influence cell tolerance to a broad range of antibiotic combinations. Analysis of the vast landscape of condition-specific drug interactions revealed a set of 24 robustly synergistic drug combinations with potential for clinical use.

A scalable, open-source implementation of a large-scale mechanistic model for single cell proliferation and death signaling

Mechanistic models of how single cells respond to different perturbations can help integrate disparate big data sets or predict response to varied drug combinations. However, the construction and simulation of such models have proved challenging. Here, we developed a python-based model creation and simulation pipeline that converts a few structured text files into an SBML standard and is high-performance- and cloud-computing ready. We applied this pipeline to our large-scale, mechanistic pan-cancer signaling model (named SPARCED) and demonstrate it by adding an IFNγ pathway submodel. We then investigated whether a putative crosstalk mechanism could be consistent with experimental observations from the LINCS MCF10A Data Cube that IFNγ acts as an anti-proliferative factor. The analyses suggested this observation can be explained by IFNγ-induced SOCS1 sequestering activated EGF receptors. This work forms a foundational recipe for increased mechanistic model-based data integration on a single-cell level, an important building block for clinically-predictive mechanistic models.

Roles of Bacterial Mechanosensitive Channels in Infection and Antibiotic Susceptibility

Bacteria accumulate osmolytes to prevent cell dehydration during hyperosmotic stress. A sudden change to a hypotonic environment leads to a rapid water influx, causing swelling of the protoplast. To prevent cell lysis through osmotic bursting, mechanosensitive channels detect changes in turgor pressure and act as emergency-release valves for the ions and osmolytes, restoring the osmotic balance. This adaptation mechanism is well-characterized with respect to the osmotic challenges bacteria face in environments such as soil or an aquatic habitat. However, mechanosensitive channels also play a role during infection, e.g., during host colonization or release into environmental reservoirs. Moreover, recent studies have proposed roles for mechanosensitive channels as determinants of antibiotic susceptibility. Interestingly, some studies suggest that they serve as entry gates for antimicrobials into cells, enhancing antibiotic efficiency, while others propose that they play a role in antibiotic-stress adaptation, reducing susceptibility to certain antimicrobials. These findings suggest different facets regarding the relevance of mechanosensitive channels during infection and antibiotic exposure as well as illustrate that they may be interesting targets for antibacterial chemotherapy. Here, we summarize the recent findings on the relevance of mechanosensitive channels for bacterial infections, including transitioning between host and environment, virulence, and susceptibility to antimicrobials, and discuss their potential as antibacterial drug targets.

Sodium dehydroacetate confers broad antibiotic tolerance by remodeling bacterial metabolism

Antibiotic tolerance has been a growing crisis that is seriously threatening global public health. However, little is known about the exogenous factors capable of triggering the development of antibiotic tolerance, particularly in vivo. Here we uncovered that an previously approved food additive termed sodium dehydroacetate (DHA-S) supplementation remarkably impaired the activity of bactericidal antibiotics against various bacterial pathogens. Mechanistic studies indicated that DHA-S induced glyoxylate shunt and reduced bacterial cellular respiration by inhibiting the enzymatic activity of α-ketoglutarate dehydrogenase (α-KGDH). Furthermore, DHA-S mitigated oxidative stress imposed by bactericidal antibiotics and enhanced the function of multidrug efflux pumps. These actions worked together to induce bacterial tolerance to antibiotic killing. Interestingly, the addition of five exogenous amino acids, particularly cysteine and proline, effectively reversed antibiotic tolerance elicited by DHA-S both in vitro and in mouse models of infection. Taken together, these findings advance our understanding of the potential risks of DHA-S in the treatment of bacterial infections, and shed new insights into the relationships between antibiotic tolerance and bacterial metabolism.

Machine learning may accelerate the recognition and control of microplastic pollution: Future prospects

Microplastics (MPs, sizes <5 mm) have been found to be widely distributed in various environments, such as marine, freshwater, terrestrial and atmospheric systems. Machine learning provides a potential solution for evaluating the ecological risks of MPs based on big data. Compared with traditional models, data-driven machine learning can accelerate the realization of the control of hazardous MPs and reduce the impact of MPs at both local and global scales. However, there are some urgent issues that should be resolved. For example, lack of MP databases and incomparable literatures causing the current MP data cannot fully support big data research. Therefore, it is imperative to formulate a set of standard and universal MP collection and testing protocols. For machine learning, predictions of large-scale MP distribution and the corresponding environmental risks remain lacking. To accelerate studies of MPs in the future, the methods and theories achieved for other particle pollutants, such as nanomaterials and aerosols, can be referenced. Beyond predication alone, the improvement of causality and interpretability of machine learning deserves attention in the studies of MP risks. Overall, this perspective paper provides insights for the development of machine learning methods in research on the environmental risks of MPs.

A broadly applicable, stress-mediated bacterial death pathway regulated by the phosphotransferase system (PTS) and the cAMP-Crp cascade

Recent work indicates that killing of bacteria by diverse antimicrobial classes can involve reactive oxygen species (ROS), as if a common, self-destructive response to antibiotics occurs. However, the ROS-bacterial death theory has been challenged. To better understand stress-mediated bacterial death, we enriched spontaneous antideath mutants of Escherichia coli that survive treatment by diverse bactericidal agents that include antibiotics, disinfectants, and environmental stressors, without a priori consideration of ROS. The mutants retained bacteriostatic susceptibility, thereby ruling out resistance. Surprisingly, pan-tolerance arose from carbohydrate metabolism deficiencies in ptsI (phosphotransferase) and cyaA (adenyl cyclase); these genes displayed the activity of upstream regulators of a widely shared, stress-mediated death pathway. The antideath effect was reversed by genetic complementation, exogenous cAMP, or a Crp variant that bypasses cAMP binding for activation. Downstream events comprised a metabolic shift from the TCA cycle to glycolysis and to the pentose phosphate pathway, suppression of stress-mediated ATP surges, and reduced accumulation of ROS. These observations reveal how upstream signals from diverse stress-mediated lesions stimulate shared, late-stage, ROS-mediated events. Cultures of these stable, pan-tolerant mutants grew normally and were therefore distinct from tolerance derived from growth defects described previously. Pan-tolerance raises the potential for unrestricted disinfectant use to contribute to antibiotic tolerance and resistance. It also weakens host defenses, because three agents (hypochlorite, hydrogen peroxide, and low pH) affected by pan-tolerance are used by the immune system to fight infections. Understanding and manipulating the PtsI-CyaA-Crp–mediated death process can help better control pathogens and maintain beneficial microbiota during antimicrobial treatment.

A genome-wide atlas of antibiotic susceptibility targets and pathways to tolerance

Detailed knowledge on how bacteria evade antibiotics and eventually develop resistance could open avenues for novel therapeutics and diagnostics. It is thereby key to develop a comprehensive genome-wide understanding of how bacteria process antibiotic stress, and how modulation of the involved processes affects their ability to overcome said stress. Here we undertake a comprehensive genetic analysis of how the human pathogen Streptococcus pneumoniae responds to 20 antibiotics. We build a genome-wide atlas of drug susceptibility determinants and generated a genetic interaction network that connects cellular processes and genes of unknown function, which we show can be used as therapeutic targets. Pathway analysis reveals a genome-wide atlas of cellular processes that can make a bacterium less susceptible, and often tolerant, in an antibiotic specific manner. Importantly, modulation of these processes confers fitness benefits during active infections under antibiotic selection. Moreover, screening of sequenced clinical isolates demonstrates that mutations in genes that decrease antibiotic sensitivity and increase tolerance readily evolve and are frequently associated with resistant strains, indicating such mutations could be harbingers for the emergence of antibiotic resistance.

Expanding the search for small-molecule antibacterials by multidimensional profiling

New techniques for systematic profiling of small-molecule effects can enhance traditional growth inhibition screens for antibiotic discovery and change how we search for new antibacterial agents. Computational models that integrate physicochemical compound properties with their phenotypic and molecular downstream effects can not only predict efficacy of molecules yet to be tested, but also reveal unprecedented insights on compound modes of action (MoAs). The unbiased characterization of compounds that themselves are not growth inhibitory but exhibit diverse MoAs, can expand antibacterial strategies beyond direct inhibition of core essential functions. Early and systematic functional annotation of compound libraries thus paves the way to new models in the selection of lead antimicrobial compounds. In this Review, we discuss how multidimensional small-molecule profiling and the ever-increasing computing power are accelerating the discovery of unconventional antibacterials capable of bypassing resistance and exploiting synergies with established antibacterial treatments and with protective host mechanisms.

Adenosine Awakens Metabolism to Enhance Growth-Independent Killing of Tolerant and Persister Bacteria across Multiple Classes of Antibiotics

Metabolic and growth arrest are primary drivers of antibiotic tolerance and persistence in clinically diverse bacterial pathogens. We recently showed that adenosine (ADO) suppresses bacterial growth under nutrient-limiting conditions. In the current study, we show that despite the growth-suppressive effect of ADO, extracellular ADO enhances antibiotic killing in both Gram-negative and Gram-positive bacteria by up to 5 orders of magnitude. The ADO-potentiated antibiotic activity is dependent on purine salvage and is paralleled with a suppression of guanosine tetraphosphate synthesis and the massive accumulation of ATP and GTP. These changes in nucleoside phosphates coincide with transient increases in rRNA transcription and proton motive force. The potentiation of antibiotic killing by ADO is manifested against bacteria grown under both aerobic and anaerobic conditions, and it is exhibited even in the absence of alternative electron acceptors such as nitrate. ADO potentiates antibiotic killing by generating proton motive force and can occur independently of an ATP synthase. Bacteria treated with an uncoupler of oxidative phosphorylation and NADH dehydrogenase-deficient bacteria are refractory to the ADO-potentiated killing, suggesting that the metabolic awakening induced by this nucleoside is intrinsically dependent on an energized membrane. In conclusion, ADO represents a novel example of metabolite-driven but growth-independent means to reverse antibiotic tolerance. Our investigations identify the purine salvage pathway as a potential target for the development of therapeutics that may improve infection clearance while reducing the emergence of antibiotic resistance. IMPORTANCE Antibiotic tolerance, which is a hallmark of persister bacteria, contributes to treatment-refractory infections and the emergence of heritable antimicrobial resistance. Drugs that reverse tolerance and persistence may become part of the arsenal to combat antimicrobial resistance. Here, we demonstrate that salvage of extracellular ADO reduces antibiotic tolerance in nutritionally stressed Escherichia coli, Salmonella enterica, and Staphylococcus aureus. ADO potentiates bacterial killing under aerobic and anaerobic conditions and takes place in bacteria lacking the ATP synthase. However, the sensitization to antibiotic killing elicited by ADO requires an intact NADH dehydrogenase, suggesting a requirement for an energized electron transport chain. ADO antagonizes antibiotic tolerance by activating ATP and GTP synthesis, promoting proton motive force and cellular respiration while simultaneously suppressing the stringent response. These investigations reveal an unprecedented role for purine salvage stimulation as a countermeasure of antibiotic tolerance and the emergence of antimicrobial resistance.

Synthetic molecular evolution of antimicrobial peptides

As we learn more about how peptide structure and activity are related, we anticipate that antimicrobial peptides will be engineered to have strong potency and distinct functions and that synthetic peptides will have new biomedical applications, such as treatments for emerging infectious diseases. As a result of the enormous number of possible amino acid sequences and the low-throughput nature of antimicrobial peptide assays, computational tools for peptide design and optimization are needed for direct experimentation toward obtaining functional sequences. Recent developments in computational tools have improved peptide design, saving labor, reagents, costs, and time. At the same time, improvements in peptide synthesis and experimental platforms continue to reduce the cost and increase the throughput of peptide-drug screening. In this review, we discuss the current methods of peptide design and engineering, including in silico methods and peptide synthesis and screening, and highlight areas of potential improvement.

Mechanisms Underlying Synergistic Killing of Polymyxin B in Combination with Cannabidiol against Acinetobacter baumannii: A Metabolomic Study

Polymyxins have resurged as the last-resort antibiotics against multidrug-resistant Acinetobacter baumannii. As reports of polymyxin resistance in A. baumannii with monotherapy have become increasingly common, combination therapy is usually the only remaining treatment option. A novel and effective strategy is to combine polymyxins with non-antibiotic drugs. This study aimed to investigate, using untargeted metabolomics, the mechanisms of antibacterial killing synergy of the combination of polymyxin B with a synthetic cannabidiol against A. baumannii ATCC 19606. The antibacterial synergy of the combination against a panel of Gram-negative pathogens (Acinetobacter baumannii, Klebsiella pneumoniae and Pseudomonas aeruginosa) was also explored using checkerboard and static time-kill assays. The polymyxin B–cannabidiol combination showed synergistic antibacterial activity in checkerboard and static time-kill assays against both polymyxin-susceptible and polymyxin-resistant isolates. The metabolomics study at 1 h demonstrated that polymyxin B monotherapy and the combination (to the greatest extent) significantly perturbed the complex interrelated metabolic pathways involved in the bacterial cell envelope biogenesis (amino sugar and nucleotide sugar metabolism, peptidoglycan, and lipopolysaccharide (LPS) biosynthesis), nucleotides (purine and pyrimidine metabolism) and peptide metabolism; notably, these pathways are key regulators of bacterial DNA and RNA biosynthesis. Intriguingly, the combination caused a major perturbation in bacterial membrane lipids (glycerophospholipids and fatty acids) compared to very minimal changes induced by monotherapies. At 4 h, polymyxin B–cannabidiol induced more pronounced effects on the abovementioned pathways compared to the minimal impact of monotherapies. This metabolomics study for the first time showed that in disorganization of the bacterial envelope formation, the DNA and RNA biosynthetic pathways were the most likely molecular mechanisms for the synergy of the combination. The study suggests the possibility of cannabidiol repositioning, in combination with polymyxins, for treatment of MDR polymyxin-resistant Gram-negative infections.

Iron-Induced Respiration Promotes Antibiotic Resistance in Actinomycete Bacteria

The bacterial response to antibiotics eliciting resistance is one of the key challenges in global health. Despite many attempts to understand intrinsic antibiotic resistance, many of the underlying mechanisms still remain elusive. In this study, we found that iron supplementation promoted antibiotic resistance in Streptomyces coelicolor. Iron-promoted resistance occurred specifically against bactericidal antibiotics, irrespective of the primary target of antibiotics. Transcriptome profiling revealed that some genes in the central metabolism and respiration were upregulated under iron-replete conditions. Iron supported the growth of S. coelicolor even under anaerobic conditions. In the presence of potassium cyanide, which reduces aerobic respiration of cells, iron still promoted respiration and antibiotic resistance. This suggests the involvement of a KCN-insensitive type of respiration in the iron effect. This phenomenon was also observed in another actinobacterium, Mycobacterium smegmatis. Taken together, these findings provide insight into a bacterial resistance strategy that mitigates the activity of bactericidal antibiotics whose efficacy accompanies oxidative damage by switching the respiration mode.

Vivarium: an interface and engine for integrative multiscale modeling in computational biology

MOTIVATION

This article introduces Vivarium-software born of the idea that it should be as easy as possible for computational biologists to define any imaginable mechanistic model, combine it with existing models and execute them together as an integrated multiscale model. Integrative multiscale modeling confronts the complexity of biology by combining heterogeneous datasets and diverse modeling strategies into unified representations. These integrated models are then run to simulate how the hypothesized mechanisms operate as a whole. But building such models has been a labor-intensive process that requires many contributors, and they are still primarily developed on a case-by-case basis with each project starting anew. New software tools that streamline the integrative modeling effort and facilitate collaboration are therefore essential for future computational biologists.

RESULTS

Vivarium is a software tool for building integrative multiscale models. It provides an interface that makes individual models into modules that can be wired together in large composite models, parallelized across multiple CPUs and run with Vivarium's discrete-event simulation engine. Vivarium's utility is demonstrated by building composite models that combine several modeling frameworks: agent-based models, ordinary differential equations, stochastic reaction systems, constraint-based models, solid-body physics and spatial diffusion. This demonstrates just the beginning of what is possible-Vivarium will be able to support future efforts that integrate many more types of models and at many more biological scales.

AVAILABILITY AND IMPLEMENTATION

The specific models, simulation pipelines and notebooks developed for this article are all available at the vivarium-notebooks repository: https://github.com/vivarium-collective/vivarium-notebooks. Vivarium-core is available at https://github.com/vivarium-collective/vivarium-core, and has been released on Python Package Index. The Vivarium Collective (https://vivarium-collective.github.io) is a repository of freely available Vivarium processes and composites, including the processes used in Section 3. Supplementary Materials provide with an extensive methodology section, with several code listings that demonstrate the basic interfaces.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Systems Biology of Gut Microbiota-Human Receptor Interactions: Toward Anti-inflammatory Probiotics

The incidence and prevalence of inflammatory disorders have increased globally, and is projected to double in the next decade. Gut microbiome-based therapeutics have shown promise in ameliorating chronic inflammation. However, they are largely experimental, context- or strain-dependent and lack a clear mechanistic basis. This hinders precision probiotics and poses significant risk, especially to individuals with pre-existing conditions. Molecules secreted by gut microbiota act as ligands to several health-relevant receptors expressed in human gut, such as the G-protein coupled receptors (GPCRs), Toll-like receptor 4 (TLR4), pregnane X receptor (PXR), and aryl hydrocarbon receptor (AhR). Among these, the human AhR expressed in different tissues exhibits anti-inflammatory effects and shows activity against a wide range of ligands produced by gut bacteria. However, different AhR ligands induce varying host responses and signaling in a tissue/organ-specific manner, which remain mostly unknown. The emerging systems biology paradigm, with its powerful in silico tool repertoire, provides opportunities for comprehensive and high-throughput strain characterization. In particular, combining metabolic models with machine learning tools can be useful to delineate tissue and ligand-specific signaling and thus their causal mechanisms in disease and health. The knowledge of such a mechanistic basis is indispensable to account for strain heterogeneity and actualize precision probiotics.

A machine learning approach-based array sensor for rapidly predicting the mechanisms of action of antibacterial compounds

Rapid and accurate identification of the mechanisms of action (MoAs) of antibacterial compounds remains a challenge for the development of antibacterial compounds. Computational inference methods for determining the MoAs of antibacterial compounds have been developed in recent years. In particular, approaches combining machine learning technology enable precisely recognizing the MoA of antibacterial compounds. However, these methods heavily rely on the big data resulting from multiplexed experiments. As such, these approaches tend to produce minimal throughput and are not comprehensive enough to be adapted to widespread industrial applications. Here, we present a machine learning approach based on a customized array sensor for directly identifying the MoAs of antibacterial compounds. The array sensor consists of different two-dimensional nanomaterial fluorescence quenchers with different fluorescence-labeled single-stranded DNAs (ssDNAs). By mapping the subtle difference of the physicochemical properties on the bacterial surface treated with different antibacterial compound stimuli, the array sensor ensures visualizing the recognition process. Moreover, the customized array sensor produces a high volume of the MoA database, overcoming the dependence on big data. We further use the array sensor to build a chemical-response unique "fingerprint" database of MoAs. By combining a neural network-based genetic algorithm (NNGA), we rapidly discriminate the MoAs of four antibiotics with an overall accuracy of 100%. Furthermore, a new screening antibacterial peptide has been discovered and evaluated by our approach for determining the MoA with high accuracy proven by other techniques.

Ceragenins and Antimicrobial Peptides Kill Bacteria through Distinct Mechanisms

Ceragenins are a family of synthetic amphipathic molecules designed to mimic the properties of naturally occurring cationic antimicrobial peptides (CAMPs). Although ceragenins have potent antimicrobial activity, whether their mode of action is similar to that of CAMPs has remained elusive. Here, we reported the results of a comparative study of the bacterial responses to two well-studied CAMPs, LL37 and colistin, and two ceragenins with related structures, CSA13 and CSA131. Using transcriptomic and proteomic analyses, we found that Escherichia coli responded similarly to both CAMPs and ceragenins by inducing a Cpx envelope stress response. However, whereas E. coli exposed to CAMPs increased expression of genes involved in colanic acid biosynthesis, bacteria exposed to ceragenins specifically modulated functions related to phosphate transport, indicating distinct mechanisms of action between these two classes of molecules. Although traditional genetic approaches failed to identify genes that confer high-level resistance to ceragenins, using a Clustered Regularly Interspaced Short Palindromic Repeats interference (CRISPRi) approach we identified E. coli essential genes that when knocked down modify sensitivity to these molecules. Comparison of the essential gene-antibiotic interactions for each of the CAMPs and ceragenins identified both overlapping and distinct dependencies for their antimicrobial activities. Overall, this study indicated that, while some bacterial responses to ceragenins overlap those induced by naturally occurring CAMPs, these synthetic molecules target the bacterial envelope using a distinctive mode of action. IMPORTANCE The development of novel antibiotics is essential because the current arsenal of antimicrobials will soon be ineffective due to the widespread occurrence of antibiotic resistance. The development of naturally occurring cationic antimicrobial peptides (CAMPs) for therapeutics to combat antibiotic resistance has been hampered by high production costs and protease sensitivity, among other factors. The ceragenins are a family of synthetic CAMP mimics that kill a broad spectrum of bacterial species but are less expensive to produce, resistant to proteolytic degradation, and seemingly resistant to the development of high-level resistance. Determining how ceragenins function may identify new essential biological pathways of bacteria that are less prone to the development of resistance and will further our understanding of the design principles for maximizing the effects of synthetic CAMPs.

Screening of antibacterial compounds with novel structure from the FDA approved drugs using machine learning methods

Bacterial infection is one of the most important factors affecting the human life span. Elderly people are more harmed by bacterial infections due to their deficits in immunity. Because of the lack of new antibiotics in recent years, bacterial resistance has increasingly become a serious problem globally. In this study, an antibacterial compound predictor was constructed using the support vector machines and random forest methods and the data of the active and inactive antibacterial compounds from the ChEMBL database. The results showed that both models have excellent prediction performance (mean accuracy >0.9 and mean AUC >0.9 for the two models). We used the predictor to screen potential antibacterial compounds from FDA-approved drugs in the DrugBank database. The screening results showed that 1087 small-molecule drugs have potential antibacterial activity and 154 of them are FDA-approved antibacterial drugs, which accounts for 76.2% of the approved antibacterial drugs collected in this study. Through molecular fingerprint similarity analysis and common substructure analysis, we screened 8 predicted antibacterial small-molecule compounds with novel structures compared with known antibacterial drugs, and 5 of them are widely used in the treatment of various tumors. This study provides a new insight for predicting antibacterial compounds by using approved drugs, the predicted compounds might be used to treat bacterial infections and extend lifespan.

Proteomics in antibiotic resistance and tolerance research: Mapping the resistome and the tolerome of bacterial pathogens

Antibiotic resistance, the ability of a microbial pathogen to evade the effects of antibiotics thereby allowing them to grow under elevated drug concentrations, is an alarming health problem worldwide and has attracted the attention of scientists for decades. On the other hand, the clinical importance of persistence and tolerance as alternative mechanisms for pathogens to survive prolonged lethal antibiotic doses has recently become increasingly appreciated. Persisters and high-tolerance populations are thought to cause the relapse of infectious diseases, and provide opportunities for the pathogens to evolve resistance during the course of antibiotic therapy. Although proteomics and other omics methodology have long been employed to study resistance, its applications in studying persistence and tolerance are still limited. However, due to the growing interest in the topic and recent progress in method developments to study them, there have been some proteomic studies that yield fresh insights into the phenomenon of persistence and tolerance. Combined with the studies on resistance, these collectively guide us to novel molecular targets for the potential drugs for the control of these dangerous pathogens. In this review, we surveyed previous proteomic studies to investigate resistance, persistence, and tolerance mechanisms, and discussed emerging experimental strategies for studying these phenotypes with a combination of adaptive laboratory evolution and high-throughput proteomics. This article is protected by copyright. All rights reserved.

The Serum and Fecal Metabolomic Profiles of Growing Kittens Treated with Amoxicillin/Clavulanic Acid or Doxycycline

Simple Summary

This study investigated the impact of antibiotic treatment οn the serum and fecal metabolome (the collection of all small molecules produced by the gut bacteria and the host) of young cats. Thirty 2-month-old cats with an upper respiratory tract infection were treated with either amoxicillin/clavulanic acid for 20 days or doxycycline for 28 days. In addition, another 15 control cats that did not receive antibiotics were included. Blood was collected on days 0 (before treatment), 20/28 (last day of treatment), and 300 (10 months after the end of treatment), while feces were collected on days 0, 20/28, 60, 120, and 300. Seven serum and fecal metabolites differed between cats treated with antibiotics and control cats at the end of treatment period. Ten months after treatment, no metabolites differed from healthy cats, suggesting that amoxicillin/clavulanic acid or doxycycline treatment only temporarily affects the abundance of the serum and fecal metabolome.

Abstract

The long-term impact of antibiotics on the serum and fecal metabolome of kittens has not yet been investigated. Therefore, the objective of this study was to evaluate the serum and fecal metabolome of kittens with an upper respiratory tract infection (URTI) before, during, and after antibiotic treatment and compare it with that of healthy control cats. Thirty 2-month-old cats with a URTI were randomly assigned to receive either amoxicillin/clavulanic acid for 20 days or doxycycline for 28 days, and 15 cats of similar age were enrolled as controls. Fecal samples were collected on days 0, 20/28, 60, 120, and 300, while serum was collected on days 0, 20/28, and 300. Untargeted and targeted metabolomic analyses were performed on both serum and fecal samples. Seven metabolites differed significantly in antibiotic-treated cats compared to controls on day 20/28, with two differing on day 60, and two on day 120. Alterations in the pattern of serum amino acids, antioxidants, purines, and pyrimidines, as well as fecal bile acids, sterols, and fatty acids, were observed in antibiotic-treated groups that were not observed in control cats. However, the alterations caused by either amoxicillin/clavulanic acid or doxycycline of the fecal and serum metabolome were only temporary and were resolved by 10 months after their withdrawal.

Metabolic signatures of regulation by phosphorylation and acetylation

Acetylation and phosphorylation are highly conserved posttranslational modifications (PTMs) that regulate cellular metabolism, yet how metabolic control is shared between these PTMs is unknown. Here we analyze transcriptome, proteome, acetylome, and phosphoproteome datasets in E. coli, S. cerevisiae, and mammalian cells across diverse conditions using CAROM, a new approach that uses genome-scale metabolic networks and machine learning to classify targets of PTMs. We built a single machine learning model that predicted targets of each PTM in a condition across all three organisms based on reaction attributes (AUC>0.8). Our model predicted phosphorylated enzymes during a mammalian cell-cycle, which we validate using phosphoproteomics. Interpreting the machine learning model using game theory uncovered enzyme properties including network connectivity, essentiality, and condition-specific factors such as maximum flux that differentiate targets of phosphorylation from acetylation. The conserved and predictable partitioning of metabolic regulation identified here between these PTMs may enable rational rewiring of regulatory circuits.

Towards a metagenomics machine learning interpretable model for understanding the transition from adenoma to colorectal cancer

Gut microbiome is gaining interest because of its links with several diseases, including colorectal cancer (CRC), as well as the possibility of being used to obtain non-intrusive predictive disease biomarkers. Here we performed a meta-analysis of 1042 fecal metagenomic samples from seven publicly available studies. We used an interpretable machine learning approach based on functional profiles, instead of the conventional taxonomic profiles, to produce a highly accurate predictor of CRC with better precision than those of previous proposals. Moreover, this approach is also able to discriminate samples with adenoma, which makes this approach very promising for CRC prevention by detecting early stages in which intervention is easier and more effective. In addition, interpretable machine learning methods allow extracting features relevant for the classification, which reveals basic molecular mechanisms accounting for the changes undergone by the microbiome functional landscape in the transition from healthy gut to adenoma and CRC conditions. Functional profiles have demonstrated superior accuracy in predicting CRC and adenoma conditions than taxonomic profiles and additionally, in a context of explainable machine learning, provide useful hints on the molecular mechanisms operating in the microbiota behind these conditions.

Integrating genome-scale metabolic modelling and transfer learning for human gene regulatory network reconstruction

MOTIVATION

Gene regulation is responsible for controlling numerous physiological functions and dynamically responding to environmental fluctuations. Reconstructing the human network of gene regulatory interactions is thus paramount to understanding the cell functional organization across cell types, as well as to elucidating pathogenic processes and identifying molecular drug targets. Although significant effort has been devoted towards this direction, existing computational methods mainly rely on gene expression levels, possibly ignoring the information conveyed by mechanistic biochemical knowledge. Moreover, except for a few recent attempts, most of the existing approaches only consider the information of the organism under analysis, without exploiting the information of related model organisms.

RESULTS

We propose a novel method for the reconstruction of the human gene regulatory network, based on a transfer learning strategy that synergically exploits information from human and mouse, conveyed by gene-related metabolic features generated in silico from gene expression data. Specifically, we learn a predictive model from metabolic activity inferred via tissue-specific metabolic modelling of artificial gene knockouts. Our experiments show that the combination of our transfer learning approach with the constructed metabolic features provides a significant advantage in terms of reconstruction accuracy, as well as additional clues on the contribution of each constructed metabolic feature.

AVAILABILITY AND IMPLEMENTATION

The method, the datasets and all the results obtained in this study are available at: https://doi.org/10.6084/m9.figshare.c.5237687.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

A Practical Guide to Integrating Multimodal Machine Learning and Metabolic Modeling

Complex, distributed, and dynamic sets of clinical biomedical data are collectively referred to as multimodal clinical data. In order to accommodate the volume and heterogeneity of such diverse data types and aid in their interpretation when they are combined with a multi-scale predictive model, machine learning is a useful tool that can be wielded to deconstruct biological complexity and extract relevant outputs. Additionally, genome-scale metabolic models (GSMMs) are one of the main frameworks striving to bridge the gap between genotype and phenotype by incorporating prior biological knowledge into mechanistic models. Consequently, the utilization of GSMMs as a foundation for the integration of multi-omic data originating from different domains is a valuable pursuit towards refining predictions. In this chapter, we show how cancer multi-omic data can be analyzed via multimodal machine learning and metabolic modeling. Firstly, we focus on the merits of adopting an integrative systems biology led approach to biomedical data mining. Following this, we propose how constraint-based metabolic models can provide a stable yet adaptable foundation for the integration of multimodal data with machine learning. Finally, we provide a step-by-step tutorial for the combination of machine learning and GSMMs, which includes: (i) tissue-specific constraint-based modeling; (ii) survival analysis using time-to-event prediction for cancer; and (iii) classification and regression approaches for multimodal machine learning. The code associated with the tutorial can be found at https://github.com/Angione-Lab/Tutorials_Combining_ML_and_GSMM .

Artificial intelligence in pathological evaluation of gastrointestinal cancers

The integration of artificial intelligence (AI) has shown promising benefits in many fields of diagnostic histopathology, including for gastrointestinal cancers (GCs), such as tumor identification, classification, and prognosis prediction. In parallel, recent evidence suggests that AI may help reduce the workload in gastrointestinal pathology by automatically detecting tumor tissues and evaluating prognostic parameters. In addition, AI seems to be an attractive tool for biomarker/genetic alteration prediction in GC, as it can contain a massive amount of information from visual data that is complex and partially understandable by pathologists. From this point of view, it is suggested that advances in AI could lead to revolutionary changes in many fields of pathology. Unfortunately, these findings do not exclude the possibility that there are still many hurdles to overcome before AI applications can be safely and effectively applied in actual pathology practice. These include a broad spectrum of challenges from needs identification to cost-effectiveness. Therefore, unlike other disciplines of medicine, no histopathology-based AI application, including in GC, has ever been approved either by a regulatory authority or approved for public reimbursement. The purpose of this review is to present data related to the applications of AI in pathology practice in GC and present the challenges that need to be overcome for their implementation.