Interpreting tree ensemble machine learning models with endoR

Background: Tree ensemble machine learning models are increasingly used in microbiome science as they are compatible with the compositional, high-dimensional, and sparse structure of sequence-based microbiome data. While such models are often good at predicting phenotypes based on microbiome data, they only yield limited insights into how microbial taxa or genomic content may be associated. Results: We developed endoR, a method to interpret a fitted tree ensemble model. First, endoR simplifies the fitted model into a decision ensemble from which it then extracts information on the importance of individual features and their pairwise interactions and also visualizes these data as an interpretable network. Both the network and importance scores derived from endoR provide insights into how features, and interactions between them, contribute to the predictive performance of the fitted model. Adjustable regularization and bootstrapping help reduce the complexity and ensure that only essential parts of the model are retained. We assessed the performance of endoR on both simulated and real metagenomic data. We found endoR to infer true associations with more or comparable accuracy than other commonly used approaches while easing and enhancing model interpretation. Using endoR, we also confirmed published results on gut microbiome differences between cirrhotic and healthy individuals. Finally, we utilized endoR to gain insights into components of the microbiome that predict the presence of human gut methanogens, as these hydrogen-consumers are expected to interact with fermenting bacteria in a complex syntrophic network. Specifically, we analyzed a global metagenome dataset of 2203 individuals and confirmed the previously reported association between Methanobacteriaceae and Christensenellales. Additionally, we observed that Methanobacteriaceae are associated with a network of hydrogen-producing bacteria. Conclusion: Our method accurately captures how tree ensembles use features and interactions between them to predict a response. As demonstrated by our applications, the resultant visualizations and summary outputs facilitate model interpretation and enable the generation of novel hypotheses about complex systems. An implementation of endoR is available as an open-source R-package on GitHub (https://github.com/leylabmpi/endoR).

Genomic Features and Construction of Streamlined Genome Chassis of Nisin Z Producer Lactococcus lactis N8

Lactococcus lactis is a commonly used fermenting bacteria in cheese, beverages and meat products. Due to the lack of simplified chassis strains, it has not been widely used in the fields of synthetic biology. Thus, the construction of lactic acid bacteria chassis strains becomes more and more important. In this study, we performed whole genome sequencing, annotation and analysis of L. lactis N8. Based on the genome analysis, we found that L. lactis N8 contains two large plasmids, and the function prediction of the plasmids shows that some regions are related to carbohydrate transport/metabolism, multi-stress resistance and amino acid uptake. L. lactis N8 contains a total of seven prophage-related fragments and twelve genomic islands. A gene cluster encoding a hybrid NRPS–PKS system that was found in L. lactis N8 reveals that the strain has the potential to synthesize novel secondary metabolites. Furthermore, we have constructed a simplified genome chassis of L. lactis N8 and achieved the largest amount of deletion of L. lactis so far. Taken together, the present study offers further insights into the function and potential role of L. lactis N8 as a model strain of lactic acid bacteria and lays the foundation for its application in the field of synthetic biology.

The Fecal Microbiota of Dogs Switching to a Raw Diet Only Partially Converges to That of Wolves

The genomic signature of dog domestication reveals adaptation to a starch-rich diet compared with their ancestor wolves. Diet is a key element to shape gut microbial populations in a direct way as well as through coevolution with the host. We investigated the dynamics in the gut microbiota of dogs when shifting from a starch-rich, processed kibble diet to a nature-like raw meat diet, using wolves as a wild reference. Six healthy wolves from a local zoo and six healthy American Staffordshire Terriers were included. Dogs were fed the same commercial kibble diet for at least 3 months before sampling at day 0 (DC), and then switched to a raw meat diet (the same diet as the wolves) for 28 days. Samples from the dogs were collected at day 1 (DR1), week 1 (DR7), 2 (DR14), 3 (DR21), and 4 (DR28). The data showed that the microbial population of dogs switched from kibble diet to raw diet shifts the gut microbiota closer to that of wolves, yet still showing distinct differences. At phylum level, raw meat consumption increased the relative abundance of Fusobacteria and Bacteroidetes at DR1, DR7, DR14, and DR21 (q < 0.05) compared with DC, whereas no differences in these two phyla were observed between DC and DR28. At genus level, Faecalibacterium, Catenibacterium, Allisonella, and Megamonas were significantly lower in dogs consuming the raw diet from the first week onward and in wolves compared with dogs on the kibble diet. Linear discriminant analysis effect size (LEfSe) showed a higher abundance of Stenotrophomonas, Faecalibacterium, Megamonas, and Lactobacillus in dogs fed kibble diet compared with dogs fed raw diet for 28 days and wolves. In addition, wolves had greater unidentified Lachnospiraceae compared with dogs irrespective of the diets. These results suggested that carbohydrate-fermenting bacteria give way to protein fermenters when the diet is shifted from kibble to raw diet. In conclusion, some microbial phyla, families, and genera in dogs showed only temporary change upon dietary shift, whereas some microbial groups moved toward the microbial profile of wolves. These findings open the discussion on the extent of coevolution of the core microbiota of dogs throughout domestication.

1,3-Diethylbenzimidazolium Triodide in Pharmacotherapy of Chronic Compensated Tonsillopharyngitis

The article presents the results of a clinical and bacteriological assessment of the pharmacological efficacy of 1,3-diethylbenzimidazolium triiodide in chronic compensated tonsillopharyngitis. As a result of a 10-day course of treatment, an improvement in the clinical status of patients was achieved, as well as a significant positive effect on the composition of the pharyngeal microbiocenosis. In particular, the content of Staphylococcus aureus, β-hemolytic streptococci, Escherichia signifi cantly decreased, and Enterobacteria, non-fermenting bacteria, and Streptococci pneumonia completely disappeared from the microbial focus. However, the conducted 10-day treatment did not lead to the complete disappearance of α-hemolytic and non-hemolytic streptococci from the microbial population. This may be explained by the reparative effect of Stellanin® on the lymphoid formations of the pharynx, contributing to the restoration of colonization resistance.

Bisphenol A Effects in Aqueous Environment on Lemna minor

The link between different plastic waste pollutants and their impact on the natural aquatic environment and food chain remains a constant and growing issue. Bisphenol A (BPA), a known endocrine disruptor produced in large quantities primarily in the industry of polycarbonate plastics, can accumulate in vegetal and animal tissue, thus magnifying through trophic levels. In this study we exposed viable specimens of the aquatic plant Lemna minor under controlled conditions to 50, 100 and 200 ppm BPA levels in order to partially observe the toxic effects of BPA. Colonies ceased to form during the exposure and chlorosis was present especially in the 100 ppm group. Interestingly enough, a high density formation of non-fermenting bacteria as well as coliforms was also observed in the BPA exposed cultures but not in the control groups. The levels of Malondialdehyde (MDA) in the vegetal tissue indicated cellular insults and severe damage, results that were correlated with the HPLC BPA determined concentrations of 0.1%, 0.2% and 0.4%.

Isolation of Lactobacillus delbrueckii subsp. bulgaricus and Streptococcus thermophilus from nature: Technological characterisation and antibiotic resistance

Yoghurt fermenting bacteria were isolated from natural sources including plants, dew, and rain samples (total of 300 samples) by the same methods nomadic peoples used for several centuries in Turkey. Inoculation into the reconstituted skim milk followed by planting on specific media and matrix-assisted laser desorption/ionization-time-of-flight mass spectrometry (MALDI-TOF MS) analysis allowed for the identification of 18 Lactobacillus delbrueckii subsp. and 26 Streptococcus thermophilus. A multiplex polymerase chain reaction (PCR) assay applied to lactobacilli enabled the identification of 5 isolates as L. delbrueckii subsp. bulgaricus. The isolates showed a varying range of acidification rates and proteolytic activity in reconstituted skimmed milk (RSM). S. thermophilus isolates showed a broader range of resistance and the most frequent resistance was observed for streptomycin (69.2%), gentamycin (65.3%), clindamycin (61.5%), ampicillin (61.5%), kanamycin (53.8%), and erythromycin (50%). For L. delbrueckii subsp. the highest resistance was determined for vancomycin (38.8%), ciprofloxacin (33.3%), and penicillin (27.8%). The frequency of multiple resistance was tested on 14 different antimicrobials determining that 19 S. thermophilus (73%) and 3 L. delbrueckii subsp. (16.7%) demonstrated resistance to more than three different antibiotics. In contrast to this wide-ranging resistance, five isolates from each genus were found to be susceptible to all tested antibiotics. The present study indicates that lactic acid bacteria (LAB) isolated from nature may have broad-range of resistance to antibiotics and could be a source for the transfer of resistance.

Using Statistical Equation Modeling to Understand Interactions Between Bacterial and Archaeal Populations and Volatile Fatty Acid Proportions in the Rumen

Microbial syntrophy (obligate metabolic mutualism) is the hallmark of energy-constrained anaerobic microbial ecosystems. For example, methanogenic archaea and fermenting bacteria coexist by interspecies hydrogen transfer in the complex microbial ecosystem in the foregut of ruminants; however, these synergistic interactions between different microbes in the rumen are seldom investigated. We hypothesized that certain bacteria and archaea interact and form specific microbial cohorts in the rumen. To this end, we examined the total (DNA-based) and potentially metabolically active (cDNA-based) bacterial and archaeal communities in rumen samples of dairy cows collected at different times in a 24 h period. Notably, we found the presence of distinct bacterial and archaeal networks showing potential metabolic interactions that were correlated with molar proportions of specific volatile fatty acids (VFAs). We employed hypothesis-driven structural equation modeling to test the significance of and to quantify the extent of these relationships between bacteria-archaea-VFAs in the rumen. Furthermore, we demonstrated that these distinct microbial networks were host-specific and differed between cows indicating a natural variation in specific microbial networks in the rumen of dairy cows. This study provides new insights on potential microbial metabolic interactions in anoxic environments that have broader applications in methane mitigation, energy conservation, and agricultural production.

Isolation and Identification of Cellulase Producing and Sugar Fermenting Bacteria for Second-Generation Bioethanol Production

Over the last decades, the negative impacts of fossil fuel on the environment and increasing demand for energy due to the unavoidable depletion of fossil fuels, has transformed the world’s interests towards alternative fuels. In particular, bioethanol production from cellulosic biomass for the transportation sector has been incrementing since the last decade. The bacterial pathway for bioethanol production is a relatively novel concept and the present study focused on the isolation of potential “cellulase-producing” bacteria from cow dung, compost soil, and termite gut and isolating sugar fermenting bacteria from palm wine. To select potential candidates for cellulase enzyme production, primary and secondary assays were conducted using the Gram’s iodine stain in Carboxy Methyl Cellulose (CMC) medium and the Dinitrosalicylic acid (DNS) assays, respectively. Durham tube assay and Solid-Phase Micro-Extraction (SPME) coupled with Gas Chromatography-Mass Spectrometry (GC-MS) was used to evaluate the sugar fermenting efficiency of the isolated bacteria. Out of 48 bacterial isolates, 27 showed cellulase activity where Nocardiopsis sp. (S-6) demonstrated the highest extracellular crude enzyme activity of endoglucanase (1.56±0.021 U) and total cellulase activity (0.93±0.012 U). The second-highest extracellular crude enzyme activity of endoglucanase (0.21±0.021 U) and total cellulase activity (0.35±0.021 U) was recorded by Bacillus sp. (T-4). Out of a total of 8 bacterial isolates, Achromobacter sp. (PW-7) was positive for sugar fermentation resulting in 3.07% of ethanol in broth medium at 48 h incubation. The results of the study revealed that Nocardiopsis sp. (S-6) had the highest cellulase enzyme activity. However, the highest ethanol percentage was achieved with by having both Bacillus sp. (T-4) and Achromobacter sp. (PW-7) for the simultaneous saccharification and fermentation (SSF) method, as compared to separate hydrolysis and fermentation (SHF) methodologies.

DNA Metabarcoding of the Leachate Microbiota from Sanitary Landfill: potential for bioremediation process

Microbial profile knowledge is essential to news alternatives and improvements in current treatments and destinations of landfill leachate that contains a variety of toxic compounds produced by municipal solid waste (MSW) disposal. Environmental DNA metabarcoding is an efficient, quick, and low-cost methodology for whole communities’ characterization. In this respect, the leachate from the Sanitary Landfill of Foz do Iguaçu City showed mixed characteristics from both acid and methanogenic phases, and 16S rDNA metabarcoding suggested the dominance of fermenting bacteria belonging to Firmicutes phylum, followed by Proteobacteria, Bacteroidetes and Synergistetes. The leachate acidogenic phase ended up being masked in the chemical and physical analyzes, however was evidenced in the metabarcoding methodology. On the other hand, no specifically methanogenic group was detected in significant abundance. To future application using culturomics approaches to bioremediation the leachate sample presented groups extensively studied, especially the Pseudomonas genus to heavy metals treatments, such as cadmium.

Do Probiotics During In-Hospital Antibiotic Treatment Prevent Colonization of Gut Microbiota With Multi-Drug-Resistant Bacteria? A Randomized Placebo-Controlled Trial Comparing Saccharomyces to a Mixture of Lactobacillus, Bifidobacterium, and Saccharomyces

Objective: Most infections with Enterobacteriaceae producing AmpC β-lactamase (AmpC)-, extended-spectrum β-lactamase (ESBL)-, and carbapenemase-producing bacteria, vancomycin-resistant Enterococcus as well as naturally resistant non-fermenting bacteria such as Pseudomonas aeruginosa, are related to a prior colonization of the gut microbiota. The objective of this study was to determine whether treatment with probiotics during an antibiotic treatment could prevent the colonization of the gut microbiota with multi-drug resistant bacteria.Method: In total, 120 patients treated for 10 days with amoxicillin-clavulanate antibiotics were included in a randomized, placebo-controlled, double-blinded trial, comparing the effects of a 30 days treatment with placebo Saccharomyces boulardii CNCM I-745® and a probiotic mixture containing Saccharomyces boulardii, Lactobacillus acidophilus NCFM, Lactobacillus paracasei Lpc-37, Bifidobacterium lactis Bl-04, and Bifidobacterium lactis Bi-07 (Bactiol duo®). Study treatment was initiated within 48 h of the antibiotic being initiated. Most of the patients included were elderly with a mean age of 78 years old with multiple comorbidities. Stools were collected at the time of inclusion in the trial, at the end of the antibiotic treatment, and the end of the study treatment. These were cultured on selective antibiotic media.Results: Treatment with the probiotic mixture led to a significant decline in colonization with Pseudomonas after antibiotic treatment from 25 to 8.3% (p = 0.041). Colonization with AmpC-producing enterobacteria was transiently increased after the antibiotic treatment (p = 0.027) and declined after the probiotic intervention (p= 0.041). No significant changes were observed in the placebo and Saccharomyces groups. Up to 2 years after the trial, no infection with ESBL-producing bacteria was observed in the probiotic mixture group.Conclusion: The association of Saccharomyces boulardii with specific strains of Lactobacillus and Bifidobacterium influences antibiotic treatment by counteracting the colonization of the colon microbiota with antibiotic-resistant pathogens.