Multivariate approach for studying interactions between environmental variables and microbial communities

Enablers and hindrances to health promotion and disease prevention practices among healthcare workers in Nelson Mandela Bay Municipality, South Africa

Health promotion (HP) and disease prevention (DP) practices among healthcare workers (HCWs) are key to achieving universal health coverage. This study identified HP and DP enablers and hindrances and compared them at different healthcare levels in Nelson Mandela Bay Municipality, South Africa. An exploratory cross-sectional study using a structured questionnaire was conducted among HCWs (n = 501) from 23 hospitals. Bivariate and multinomial regression were used to analyze the data. The highest number of participants (70.46%; n = 353) were from tertiary hospitals. Thirteen and Eight categories of enablers and hindrances respectively were identified. Of these, eleven enablers and six hindrances of HP and DP were associated with tertiary hospitals; no enabler was identified at both primary and secondary while one hindrance was associated with primary level of health care. Collaboration among disciplines and organizations (Coeff: 2.16, 95% CI: 1.28-3.66) and programme planning (Coeff: 0.375, 95% CI: 0.23-0.62) were the predictors of HP and DP among medical doctors, while staff induction training (Coeff: 0.62, 95% CI: 0.40-0.95) and performance appraisal (Coeff: 1.86, 95% CI: 1.16-2.98) were the enablers among allied health workers. On the other hand, 'facility promoting treatment more than prevention' (Coeff: 2.03, 95% CI: 1.30-3.14) and 'practice guidelines incorporating HP' (Coeff: 2.79, 95% CI: 1.66-4.70) were the predictors of HP and DP hindrances among medical doctors and allied health workers respectively. Our work indicates the need for an operational strategy designed considering enabling and hindering factors to HP and DP practices for empowering HCWs and enhancing health outcomes.

Correlation and association analyses in microbiome study integrating multiomics in health and disease

Correlation and association analyses are one of the most widely used statistical methods in research fields, including microbiome and integrative multiomics studies. Correlation and association have two implications: dependence and co-occurrence. Microbiome data are structured as phylogenetic tree and have several unique characteristics, including high dimensionality, compositionality, sparsity with excess zeros, and heterogeneity. These unique characteristics cause several statistical issues when analyzing microbiome data and integrating multiomics data, such as large p and small n, dependency, overdispersion, and zero-inflation. In microbiome research, on the one hand, classic correlation and association methods are still applied in real studies and used for the development of new methods; on the other hand, new methods have been developed to target statistical issues arising from unique characteristics of microbiome data. Here, we first provide a comprehensive view of classic and newly developed univariate correlation and association-based methods. We discuss the appropriateness and limitations of using classic methods and demonstrate how the newly developed methods mitigate the issues of microbiome data. Second, we emphasize that concepts of correlation and association analyses have been shifted by introducing network analysis, microbe-metabolite interactions, functional analysis, etc. Third, we introduce multivariate correlation and association-based methods, which are organized by the categories of exploratory, interpretive, and discriminatory analyses and classification methods. Fourth, we focus on the hypothesis testing of univariate and multivariate regression-based association methods, including alpha and beta diversities-based, count-based, and relative abundance (or compositional)-based association analyses. We demonstrate the characteristics and limitations of each approaches. Fifth, we introduce two specific microbiome-based methods: phylogenetic tree-based association analysis and testing for survival outcomes. Sixth, we provide an overall view of longitudinal methods in analysis of microbiome and omics data, which cover standard, static, regression-based time series methods, principal trend analysis, and newly developed univariate overdispersed and zero-inflated as well as multivariate distance/kernel-based longitudinal models. Finally, we comment on current association analysis and future direction of association analysis in microbiome and multiomics studies.

Defining the relationship between vaginal and urinary microbiomes

BACKGROUND

Although the vaginal and urinary microbiomes have been increasingly well-characterized in health and disease, few have described the relationship between these neighboring environments. Elucidating this relationship has implications for understanding how manipulation of the vaginal microbiome may affect the urinary microbiome and treatment of common urinary conditions.

OBJECTIVE

To describe the relationship between urinary and vaginal microbiomes using 16S rRNA gene sequencing. We hypothesized that the composition of the urinary and vaginal microbiomes would be significantly associated, with similarities in predominant taxa.

STUDY DESIGN

This multicenter study collected vaginal swabs and catheterized urine samples from 186 women with mixed urinary incontinence enrolled in a parent study and 84 similarly aged controls. Investigators decided a priori that if vaginal and/or urinary microbiomes differed between continent and incontinent women, the groups would be analyzed separately; if similar, samples from continent and incontinent women would be pooled and analyzed together. A central laboratory sequenced variable regions 1-3 (v1-3) and characterized bacteria to the genus level. Operational taxonomic unit abundance was described for paired vaginal and urine samples. Pearson's correlation characterized the relationship between individual operational taxonomic units of paired samples. Canonical correlation analysis evaluated the association between clinical variables (including mixed urinary incontinence and control status) and vaginal and urinary operational taxonomic units, using the Canonical correlation analysis function in the Vegan package (R version 3.5). Linear discriminant analysis effect size was used to find taxa that discriminated between vaginal and urinary samples.

RESULTS

Urinary and vaginal samples were collected from 212 women (mean age 53±11 years) and results from 197 paired samples were available for analysis. As operational taxonomic units in mixed urinary incontinence and control samples were related in canonical correlation analysis and since taxa did not discriminate between mixed urinary incontinence or controls in either vagina or urine, mixed urinary incontinence and control samples were pooled for further analysis. Canonical correlation analysis of vaginal and urinary samples indicated that that 60 of the 100 most abundant operational taxonomic units in the samples largely overlapped. Lactobacillus was the most abundant genus in both urine and vagina (contributing on average 53% to an individual's urine sample and 64% to an individual's vaginal sample) (Pearson correlation r=0.53). Although less abundant than Lactobacillus, other bacteria with high Pearson correlation coefficients also commonly found in vagina and urine included: Gardnerella (r=0.70), Prevotella (r=0.64), and Ureaplasma (r=0.50). Linear discriminant analysis effect size analysis identified Tepidimonas and Flavobacterium as bacteria that distinguished the urinary environment for both mixed urinary incontinence and controls as these bacteria were absent in the vagina (Tepidimonas effect size 2.38, P<.001, Flavobacterium effect size 2.15, P<.001). Although Lactobacillus was the most abundant bacteria in both urine and vagina, it was more abundant in the vagina (linear discriminant analysis effect size effect size 2.72, P<.001).

CONCLUSION

Significant associations between vaginal and urinary microbiomes were demonstrated, with Lactobacillus being predominant in both urine and vagina. Abundance of other bacteria also correlated highly between the vagina and urine. This inter-relatedness has implications for studying manipulation of the urogenital microbiome in treating conditions such as urgency urinary incontinence and urinary tract infections.

Analyzing Metabolomics Data for Environmental Health and Exposome Research

The exposome is the cumulative measure of environmental influences and associated biological responses across the life span, with critical relevance for understanding how exposures can impact human health. Metabolomics analysis of biological samples offers unique advantages for examining the exposome. Simultaneous analysis of external exposures, biological responses, and host susceptibility at a systems level can help establish links between external exposures and health outcomes. As metabolomics technologies continue to evolve for the study of the exposome, metabolomics ultimately will help provide valuable insights for exposure risk assessment, and disease prevention and management. Here, we discuss recent advances in metabolomics, and describe data processing protocols that can enable analysis of the exposome. This chapter focuses on using liquid chromatography-mass spectrometry (LC-MS)-based untargeted metabolomics for analysis of the exposome, including (1) preprocessing of untargeted metabolomics data, (2) identification of exposure chemicals and their metabolites, and (3) methods to establish associations between exposures and diseases.

Mapping the microbial interactome: Statistical and experimental approaches for microbiome network inference

IMPACT STATEMENT

This review provides a comprehensive description of experimental and statistical tools used for network analyses of the human gut microbiome. Understanding the system dynamics of microbial interactions may lead to the improvement of therapeutic approaches for managing microbiome-associated diseases. Microbiome network inference tools have been developed and applied to both cross-sectional and longitudinal experimental designs, as well as to multi-omic datasets, with the goal of untangling the complex web of microbe-host, microbe-environmental, and metabolism-mediated microbial interactions. The characterization of these interaction networks may lead to a better understanding of the systems dynamics of the human gut microbiome, augmenting our knowledge of the microbiome's role in human health, and guiding the optimization of effective, precise, and rational therapeutic strategies for managing microbiome-associated disease.

Responses of Salt Marsh Plant Rhizosphere Diazotroph Assemblages to Drought

Drought has many consequences in the tidally dominated Spartina sp. salt marshes of the southeastern US; including major dieback events, changes in sediment chemistry and obvious changes in the landscape. These coastal systems tend to be highly productive, yet many salt marshes are also nitrogen limited and depend on plant associated diazotrophs as their source of ‘new’ nitrogen. A 4-year study was conducted to investigate the structure and composition of the rhizosphere diazotroph assemblages associated with 5 distinct plant zones in one such salt marsh. A period of greatly restricted tidal inundation and precipitation, as well as two periods of drought (June–July 2004, and May 2007) occurred during the study. DGGE of nifH PCR amplicons from rhizosphere samples, Principal Components Analysis of the resulting banding patterns, and unconstrained ordination analysis of taxonomic data and environmental parameters were conducted. Diazotroph assemblages were organized into 5 distinct groups (R² = 0.41, p value < 0.001) whose presence varied with the environmental conditions of the marsh. Diazotroph assemblage group detection differed during and after the drought event, indicating that persistent diazotrophs maintained populations that provided reduced supplies of new nitrogen for vegetation during the periods of drought.

Nutritional composition of honey bee food stores vary with floral composition

Sufficiently diverse and abundant resources are essential for generalist consumers, and form an important part of a suite of conservation strategies for pollinators. Honey bees are generalist foragers and are dependent on diverse forage to adequately meet their nutritional needs. Through analysis of stored pollen (bee bread) samples obtained from 26 honey bee (Apis mellifera L.) hives across NW-England, we quantified bee bread nutritional content and the plant species that produced these stores from pollen. Protein was the most abundant nutrient by mass (63%), followed by carbohydrates (26%). Protein and lipid content (but not carbohydrate) contributed significantly to ordinations of floral diversity, linking dietary quality with forage composition. DNA sequencing of the ITS2 region of the nuclear ribosomal DNA gene identified pollen from 89 distinct plant genera, with each bee bread sample containing between 6 and 35 pollen types. Dominant genera included dandelion (Taraxacum), which was positively correlated with bee bread protein content, and cherry (Prunus), which was negatively correlated with the amount of protein. In addition, proportions of amino acids (e.g. histidine and valine) varied as a function of floral species composition. These results also quantify the effects of individual plant genera on the nutrition of honey bees. We conclude that pollens of different plants act synergistically to influence host nutrition; the pollen diversity of bee bread is linked to its nutrient content. Diverse environments compensate for the loss of individual forage plants, and diversity loss may, therefore, destabilize consumer communities due to restricted access to alternative resources.

Microbes in Infant Gut Development: Placing Abundance Within Environmental, Clinical and Growth Parameters

Sound and timely microbial gut colonization completes newborn’s healthy metabolic programming and manifests in infant appropriate growth and weight development. Feces, collected at 3, 30, and 90 days after birth from 60 breastfed Slovenian newborns, was submitted to microbial DNA extraction and qPCR quantification of selected gut associated taxa. Multivariate regression analysis was applied to evaluate microbial dynamics with respect to infant demographic, environmental, clinical characteristics and first year growth data. Early microbial variability was marked by the proportion of Bacilli, but diminished and converged in later samples, as bifidobacteria started to prevail. The first month proportions of enterococci were associated with maternity hospital locality and supplementation of breastfeeding with formulae, while Enterococcus faecalis proportion reflected the mode of delivery. Group Bacteroides-Prevotella proportion was associated with infant weight and ponderal index at first month. Infant mixed feeding pattern and health issues within the first month revealed the most profound and extended microbial perturbations. Our findings raise concerns over the ability of the early feeding supplementation to emulate and support the gut microbiota in a way similar to the exclusively breastfed infants. Additionally, practicing supplementation beyond the first month also manifested in higher first year weight and weight gain Z-score.

Unboxing cluster heatmaps

Background

Cluster heatmaps are commonly used in biology and related fields to reveal hierarchical clusters in data matrices. This visualization technique has high data density and reveal clusters better than unordered heatmaps alone. However, cluster heatmaps have known issues making them both time consuming to use and prone to error. We hypothesize that visualization techniques without the rigid grid constraint of cluster heatmaps will perform better at clustering-related tasks.

Results

We developed an approach to “unbox” the heatmap values and embed them directly in the hierarchical clustering results, allowing us to use standard hierarchical visualization techniques as alternatives to cluster heatmaps. We then tested our hypothesis by conducting a survey of 45 practitioners to determine how cluster heatmaps are used, prototyping alternatives to cluster heatmaps using pair analytics with a computational biologist, and evaluating those alternatives with hour-long interviews of 5 practitioners and an Amazon Mechanical Turk user study with approximately 200 participants. We found statistically significant performance differences for most clustering-related tasks, and in the number of perceived visual clusters. Visit git.io/vw0t3 for our results.

Conclusions

The optimal technique varied by task. However, gapmaps were preferred by the interviewed practitioners and outperformed or performed as well as cluster heatmaps for clustering-related tasks. Gapmaps are similar to cluster heatmaps, but relax the heatmap grid constraints by introducing gaps between rows and/or columns that are not closely clustered. Based on these results, we recommend users adopt gapmaps as an alternative to cluster heatmaps.

Microbiome: Should we diversify from diversity?

Studies on microbiome diversity are flooding the current literature, yet lessons from ecology clearly demonstrate that diversity is just one factor to consider when analyzing an ecosystem, along with its stability, structure and function. Measures of diversity may be a useful tool for interpreting metagenomic data but the question remains as to how informative they are and what insight they may provide into the state of the microbiome. A study utilizing mathematical modeling to investigate the ecological dynamics of microbial communities has shown that diversity and stability may not always be concomitant. This finding is pertinent to the gut microbiome field, especially since diversity comparisons between healthy and pathological states frequently yield contradictory results. There is a need to broaden our approach to the analysis of microbiome data if we are to better understand this complex ecological community and its role in human health and disease.

Association of sputum microbiota profiles with severity of community-acquired pneumonia in children

BACKGROUND

Competitive interactions among bacteria in the respiratory tract microbiota influence which species can colonize and potentially contribute to pathogenesis of community-acquired pneumonia (CAP). However, understanding of the role of respiratory tract microbiota in the clinical course of pediatric CAP is limited.

METHODS

We sought to compare microbiota profiles in induced sputum and nasopharyngeal/oropharyngeal (NP/OP) samples from children and to identify microbiota profiles associated with CAP severity. We used 16S ribosomal RNA sequencing and several measures of microbiota profiles, including principal component analysis (PCA), to describe the respiratory microbiota in 383 children, 6 months to <18 years, hospitalized with CAP. We examined associations between induced sputum and NP/OP microbiota profiles and CAP severity (hospital length of stay and intensive care unit admission) using logistic regression.

RESULTS

Relative abundance of bacterial taxa differed in induced sputum and NP/OP samples. In children 6 months to < 5 years, the sputum PCA factor with high relative abundance of Actinomyces, Veillonella, Rothia, and Lactobacillales was associated with decreased odds of length of stay ≥ 4 days [adjusted odds ratio (aOR) 0.69; 95 % confidence interval (CI) 0.48-0.99]. The sputum factor with high relative abundance of Haemophilus and Pasteurellaceae was associated with increased odds of intensive care unit admission [aOR 1.52; 95 % CI 1.02-2.26]. In children 5 to < 18 years, the sputum factor with high relative abundance of Porphyromonadaceae, Bacteriodales, Lactobacillales, and Prevotella was associated with increased odds of length of stay ≥ 4 days [aOR 1.52; 95 % CI 1.02-2.26]. Taxa in NP/OP samples were not associated with CAP severity.

CONCLUSION

Certain taxa in the respiratory microbiota, which were detected in induced sputum samples, are associated with the clinical course of CAP.

Application of multivariate statistical techniques in microbial ecology

Recent advances in high-throughput methods of molecular analyses have led to an explosion of studies generating large-scale ecological data sets. In particular, noticeable effect has been attained in the field of microbial ecology, where new experimental approaches provided in-depth assessments of the composition, functions and dynamic changes of complex microbial communities. Because even a single high-throughput experiment produces large amount of data, powerful statistical techniques of multivariate analysis are well suited to analyse and interpret these data sets. Many different multivariate techniques are available, and often it is not clear which method should be applied to a particular data set. In this review, we describe and compare the most widely used multivariate statistical techniques including exploratory, interpretive and discriminatory procedures. We consider several important limitations and assumptions of these methods, and we present examples of how these approaches have been utilized in recent studies to provide insight into the ecology of the microbial world. Finally, we offer suggestions for the selection of appropriate methods based on the research question and data set structure.

Dysbiosis of upper respiratory tract microbiota in elderly pneumonia patients

Bacterial pneumonia is a major cause of morbidity and mortality in elderly. We hypothesize that dysbiosis between regular residents of the upper respiratory tract (URT) microbiome, that is balance between commensals and potential pathogens, is involved in pathogen overgrowth and consequently disease. We compared oropharyngeal microbiota of elderly pneumonia patients (n=100) with healthy elderly (n=91) by 16S-rRNA-based sequencing and verified our findings in young adult pneumonia patients (n=27) and young healthy adults (n=187). Microbiota profiles differed significantly between elderly pneumonia patients and healthy elderly (PERMANOVA, P<0.0005). Highly similar differences were observed between microbiota profiles of young adult pneumonia patients and their healthy controls. Clustering resulted in 11 (sub)clusters including 95% (386/405) of samples. We observed three microbiota profiles strongly associated with pneumonia (P<0.05) and either dominated by lactobacilli (n=11), Rothia (n=51) or Streptococcus (pseudo)pneumoniae (n=42). In contrast, three other microbiota clusters (in total n=183) were correlated with health (P<0.05) and were all characterized by more diverse profiles containing higher abundances of especially Prevotella melaninogenica, Veillonella and Leptotrichia. For the remaining clusters (n=99), the association with health or disease was less clear. A decision tree model based on the relative abundance of five bacterial community members in URT microbiota showed high specificity of 95% and sensitivity of 84% (89% and 73%, respectively, after cross-validation) for differentiating pneumonia patients from healthy individuals. These results suggest that pneumonia in elderly and young adults is associated with dysbiosis of the URT microbiome with bacterial overgrowth of single species and absence of distinct anaerobic bacteria. Whether the observed microbiome changes are a cause or a consequence of the development of pneumonia or merely coincide with disease status remains a question for future research.

The impact of breastfeeding on nasopharyngeal microbial communities in infants

RATIONALE

Breastfeeding elicits significant protection against respiratory tract infections in infancy. Modulation of respiratory microbiota might be part of the natural mechanisms of protection against respiratory diseases induced by breastfeeding.

OBJECTIVES

To study the association between breastfeeding and nasopharyngeal microbial communities, including all cultivable and noncultivable bacteria.

METHODS

In this observational study, we analyzed the microbiota of infants that had received exclusive breastfeeding (n = 101) and exclusive formula feeding (n = 101) at age 6 weeks and 6 months by 16S-based GS-FLX-titanium-pyrosequencing.

MEASUREMENTS AND MAIN RESULTS

At 6 weeks of age the overall bacterial community composition was significantly different between breastfed and formula-fed children (nonmetric multidimensional scaling, P = 0.001). Breastfed children showed increased presence and abundance of the lactic acid bacterium Dolosigranulum (relative effect size [RES], 2.61; P = 0.005) and Corynebacterium (RES, 1.98; P = 0.039) and decreased abundance of Staphylococcus (RES, 0.48; P 0.03) and anaerobic bacteria, such as Prevotella (RES, 0.25; P < 0.001) and Veillonella (RES, 0.33; P < 0.001). Predominance (>50% of the microbial profile) of Corynebacterium and Dolosigranulum was observed in 45 (44.6%) breastfed infants compared with 19 (18.8%) formula-fed infants (relative risk, 2.37; P = 0.006). Dolosigranulum abundance was inversely associated with consecutive symptoms of wheezing and number of mild respiratory tract infections experienced. At 6 months of age associations between breastfeeding and nasopharyngeal microbiota composition had disappeared.

CONCLUSIONS

Our data suggest a strong association between breastfeeding and microbial community composition in the upper respiratory tract of 6-week-old infants. Observed differences in microbial community profile may contribute to the protective effect of breastfeeding on respiratory infections and wheezing in early infancy. Clinical trial registered with www.clinicaltrials.gov (NCT 00189020).