From learning taxonomies to phylogenetic learning: integration of 16S rRNA gene data into FAME-based bacterial classification

Bacterial species identification using MALDI-TOF mass spectrometry and machine learning techniques: A large-scale benchmarking study

Today machine learning methods are commonly deployed for bacterial species identification using MALDI-TOF mass spectrometry data. However, most of the studies reported in literature only consider very traditional machine learning methods on small datasets that contain a limited number of species. In this paper we present benchmarking results on an unprecedented scale for a wide range of machine learning methods, using datasets that contain almost 100,000 spectra and more than 1000 different species. The size and the diversity of the data allow to compare three important identification scenarios that are often not distinguished in literature, i.e., identification for novel biological replicates, novel strains and novel species that are not present in the training data. The results demonstrate that in all three scenarios acceptable identification rates are obtained, but the numbers are typically lower than those reported in studies with a more limited analysis. Using hierarchical classification methods, we also demonstrate that taxonomic information is in general not well preserved in MALDI-TOF mass spectrometry data. For the novel species scenario, we apply for the first time neural networks with Monte Carlo dropout, which have shown to be successful in other domains, such as computer vision, for the detection of novel species.

Novel biotechnological approaches for monitoring and immunization against resistant to antibiotics Escherichia coli and other pathogenic bacteria

The application of next-generation molecular, biochemical and immunological methods for developing new vaccines, antimicrobial compounds, probiotics and prebiotics for zoonotic infection control has been fundamental to the understanding and preservation of the symbiotic relationship between animals and humans. With increasing rates of antibiotic use, resistant bacterial infections have become more difficult to diagnose, treat, and eradicate, thereby elevating the importance of surveillance and prevention programs. Effective surveillance relies on the availability of rapid, cost-effective methods to monitor pathogenic bacterial isolates. In this opinion article, we summarize the results of some research program initiatives for the improvement of live vaccines against avian enterotoxigenic Escherichia coli using virulence factor gene deletion and engineered vaccine vectors based on probiotics. We also describe methods for the detection of pathogenic bacterial strains in eco-environmental headspace and aerosols, as well as samples of animal and human breath, based on the composition of volatile organic compounds and fatty acid methyl esters. We explain how the introduction of these low-cost biotechnologies and protocols will provide the opportunity to enhance co-operation between networks of resistance surveillance programs and integrated routine workflows of veterinary and clinical public health microbiology laboratories.

Molecular identification and diversity analysis of dental bacteria in diabetic and non-diabetic females from Saudi Arabia

Periodontal disease is a chronic infectious disease, which is characterized by the damaged dental hard tissue by lactic acid generated by microorganisms after the fermentation of carbohydrates rich diet. The risk of periodontal disease is known to be higher in diabetic patients. We compared the diversity of five commonly occurring dental bacteria including Porphyromonas gingivalis, Tannerella forsythia, Capnocytophaga ochracea, Prevotella intermedia, and Aggregatibacter actinomycetemcomitans in 14 type-2 diabetic patients and equal numbers of healthy controls. The subgingival samples were collected using sterile paper points. We used 16S rRNA sequence specific primers for PCR-based identification of dental bacteria. Our results showed that A. actinomycetemcomitans was completely absent in control subjects but present in 43% of diabetic patients. C. ochracea was highly prevalent in diabetic patients (100%) as compared to controls (28.5%). The frequency of other three bacterial species was also higher in diabetic patients than control subjects. These findings indicate that dental bacteria are highly prevalent in subgingival pockets of diabetic patients. Therefore, proper monitoring of diabetic patients for dental care is important to prevent bacterial growth and its sequela in risky individuals. Further case-control studies using larger sample size would help in validating the association between oral diseases and diabetes.

Translational utility of a hierarchical classification strategy in biomolecular data analytics

Hierarchical classification (HC) stratifies and classifies data from broad classes into more specific classes. Unlike commonly used data classification strategies, this enables the probabilistic prediction of unknown classes at different levels, minimizing the burden of incomplete databases. Despite these advantages, its translational application in biomedical sciences has been limited. We describe and demonstrate the implementation of a HC approach for "omics-driven" classification of 15 bacterial species at various taxonomic levels achieving 90-100% accuracy, and 9 cancer types into morphological types and 35 subtypes with 99% and 76% accuracy, respectively. Unknown bacterial species were probabilistically assigned with 100% accuracy to their respective genus or family using mass spectra (n = 284). Cancer types were predicted by mRNA data (n = 1960) for most subtypes with 95-100% accuracy. This has high relevance in clinical practice where complete datasets are difficult to compile with the continuous evolution of diseases and emergence of new strains, yet prediction of unknown classes, such as bacterial species, at upper hierarchy levels may be sufficient to initiate antimicrobial therapy. The algorithms presented here can be directly translated into clinical-use with any quantitative data, and have broad application potential, from unlabeled sample identification, to hierarchical feature selection, and discovery of new taxonomic variants.

Lineage structure of Streptococcus pneumoniae may be driven by immune selection on the groEL heat-shock protein

Populations of Streptococcus pneumoniae (SP) are typically structured into groups of closely related organisms or lineages, but it is not clear whether they are maintained by selection or neutral processes. Here, we attempt to address this question by applying a machine learning technique to SP whole genomes. Our results indicate that lineages evolved through immune selection on the groEL chaperone protein. The groEL protein is part of the groESL operon and enables a large range of proteins to fold correctly within the physical environment of the nasopharynx, thereby explaining why lineage structure is so stable within SP despite high levels of genetic transfer. SP is also antigenically diverse, exhibiting a variety of distinct capsular serotypes. Associations exist between lineage and capsular serotype but these can be easily perturbed, such as by vaccination. Overall, our analyses indicate that the evolution of SP can be conceptualized as the rearrangement of modular functional units occurring on several different timescales under different pressures: some patterns have locked in early (such as the epistatic interactions between groESL and a constellation of other genes) and preserve the differentiation of lineages, while others (such as the associations between capsular serotype and lineage) remain in continuous flux.

Degradation study of lindane by novel strains Kocuria sp. DAB-1Y and Staphylococcus sp. DAB-1W

BACKGROUND

This study was carried out to isolate and characterize the bacterial strains from lindane-contaminated soil and they were also assessed for their lindane-degrading potential.

METHODS

In this study the enrichment culture method was used for isolation of  lindane degrading bacterial isolates, in which the mineral salt medium (MSM) supplemented with different concentrations of lindane was used. Further, the screening for the potential lindane degrading isolates was done using the spray plate method and colorimetric dechlorinase enzyme assay. The selected isolates were also studied for their growth response under varying range of temperature, pH, and NaCl. The finally selected isolates DAB-1Y and DAB-1W showing best lindane degradation activity was further subjected to biochemical characterization, microscopy, degradation/kinetic study, and 16S rDNA sequencing. The strain identification were performed using the biochemical characterization, microscopy and the species identifies by 16S rDNA sequence of the two isolates using the standard 16S primers, the 16 S rRNA partial sequence was analyzed through BLAST analysis and phylogenetic tree was generated based on UGPMA clustering method using MEGA7 software. This shows the phylogenetic relationship with the related strains. The two isolates of this study were finally characterized as Kocuria sp. DAB-1Y and Staphylococcus sp. DAB-1W, and their 16S rRNA sequence was submitted to GenBank database with accession numbers, KJ811539 and KX986577, respectively.

RESULTS

Out of the 20 isolates, the isolates DAB-1Y and DAB-1W exhibited best lindane-degrading activity of 94 and 98%, respectively, recorded after 8 days of incubation. The optimum growth was observed at temperature 30 °C, pH 7, and 5% NaCl observed for both isolates. Of the four isomers of hexachlorocyclohexane, isomer α and γ were the fastest degrading isomers, which were degraded up to 86 and 94% by isolates DAB-1Y and up to 93 and 98% by DAB-1W, respectively, reported after 8 days incubation. Isomer β was highly recalcitrant in which maximum 35 and 32% lindane degradation was observed even after 28 days incubation by isolates, DAB-1Y and DAB-1W, respectively. At lower lindane concentrations (1-10 mg/L), specific growth rate increased with increase in lindane concentration, maximum being 0.008 and 0.006/day for DAB-1Y and DAB-1W, respectively. The 16 S rRNA partial sequence of isolate DAB-1Y showed similarity with Kocuria sp. by BLAST analysis and was named as Kocuria sp. DAB-1Y and DAB-IW with Staphylococcus sp. DAB-1W. The 16S rDNA sequence of isolate DAB-1Y and DAB-1W was submitted to online at National Centre of Biotechnology Information (NCBI) with GenBank accession numbers, KJ811539 and KX986577, respectively.

CONCLUSIONS

This study has demonstrated that Kocuria sp. DAB-1Y and Staphylococcus sp. DAB-1W were found efficient in bioremediation of gamma-HCH and can be utilized further for biodegradation of environmental contamination of lindane and can be utilized in bioremediation program.

Visualized analysis of cellular fatty acid profiles of Vibrio parahaemolyticus strains under cold stress

Vibrio parahaemolyticus is a common foodborne bacterial pathogen, which survives in cold environments and is sometimes difficult to culture. Fatty acid analysis under cold stress was conducted for several V. parahaemolyticus strains using gas chromatography/mass spectrometry, and the results were compared with those of the controls. All the fatty acid profiles obtained were visualized by multidimensional scaling (MDS) and self-organized map (SOM). It was observed that the fatty acid profiles of V. parahaemolyticus substantially changed under cold stress. The percentage of methyl palmitate remarkably decreased and that of methyl palmitoleate (except for two strains) and methyl oleate increased. These findings demonstrate the role of fatty acids in cold stress. The changes in the fatty acid profiles illustrated by MDS and SOM could differentiate strains under cold stress from the controls and can potentially lead to a method of detecting injured cold-stressed V. parahaemolyticus.

The use of FAME analyses to discriminate between different strains of Geotrichum klebahnii with different viabilities

A considerable decline in viability of spray dried cells of Geotrichum klebahnii was observed and was attributed to an undefined alteration of the used strain. As common techniques were not able to distinguish the altered from the still viable strains, we used the fatty acid methyl ester (FAME) analysis. On the basis of FAME data we were able to discriminate the three strains under investigation. Especially the ratios of cis/trans fatty acid ratios and of saturated/unsaturated fatty acid were significantly reduced in the less viable strain, pointing to an increased stress level in this strain. These findings clearly show the applicability of the FAME analysis to detect strain alterations and that this method is therefore a suitable, fast and feasible tool for quality assurance.