Dirichlet‐multinomial modelling outperforms alternatives for analysis of microbiome and other ecological count data

A new approach to describe the taxonomic structure of microbiome and its application to assess the relationship between microbial niches

BACKGROUND

Data from microbiomes from multiple niches is often collected, but methods to analyse these often ignore associations between niches. One interesting case is that of the oral microbiome. Its composition is receiving increasing attention due to reports on its associations with general health. While the oral cavity includes different niches, multi-niche microbiome data analysis is conducted using a single niche at a time and, therefore, ignores other niches that could act as confounding variables. Understanding the interaction between niches would assist interpretation of the results, and help improve our understanding of multi-niche microbiomes.

METHODS

In this study, we used a machine learning technique called latent Dirichlet allocation (LDA) on two microbiome datasets consisting of several niches. LDA was used on both individual niches and all niches simultaneously. On individual niches, LDA was used to decompose each niche into bacterial sub-communities unveiling their taxonomic structure. These sub-communities were then used to assess the relationship between microbial niches using the global test. On all niches simultaneously, LDA allowed us to extract meaningful microbial patterns. Sets of co-occurring operational taxonomic units (OTUs) comprising those patterns were then used to predict the original location of each sample.

RESULTS

Our approach showed that the per-niche sub-communities displayed a strong association between supragingival plaque and saliva, as well as between the anterior and posterior tongue. In addition, the LDA-derived microbial signatures were able to predict the original sample niche illustrating the meaningfulness of our sub-communities. For the multi-niche oral microbiome dataset we had an overall accuracy of 76%, and per-niche sensitivity of up to 83%. Finally, for a second multi-niche microbiome dataset from the entire body, microbial niches from the oral cavity displayed stronger associations to each other than with those from other parts of the body, such as niches within the vagina and the skin.

CONCLUSION

Our LDA-based approach produces sets of co-occurring taxa that can describe niche composition. LDA-derived microbial signatures can also be instrumental in summarizing microbiome data, for both descriptions as well as prediction.

ATM-deficiency-induced microglial activation promotes neurodegeneration in ataxia-telangiectasia

While ATM loss of function has long been identified as the genetic cause of ataxia-telangiectasia (A-T), how it leads to selective and progressive degeneration of cerebellar Purkinje and granule neurons remains unclear. ATM expression is enriched in microglia throughout cerebellar development and adulthood. Here, we find evidence of microglial inflammation in the cerebellum of patients with A-T using single-nucleus RNA sequencing. Pseudotime analysis revealed that activation of A-T microglia preceded upregulation of apoptosis-related genes in granule and Purkinje neurons and that microglia exhibited increased neurotoxic cytokine signaling to granule and Purkinje neurons in A-T. To confirm these findings experimentally, we performed transcriptomic profiling of A-T induced pluripotent stem cell (iPSC)-derived microglia, which revealed cell-intrinsic microglial activation of cytokine production and innate immune response pathways compared to controls. Furthermore, A-T microglia co-culture with either control or A-T iPSC-derived neurons was sufficient to induce cytotoxicity. Taken together, these studies reveal that cell-intrinsic microglial activation may promote neurodegeneration in A-T.

LMdist: Local Manifold distance accurately measures beta diversity in ecological gradients

MOTIVATION

Differentiating ecosystems poses a complex, high-dimensional problem constrained by capturing relevant variation across species profiles. Researchers use pairwise distances and subsequent dimensionality reduction to highlight variation in a few dimensions. Despite popularity in analysis of ecological data, these low-dimensional visualizations can contain geometric abnormalities such as "arch" and "horseshoe" effects, potentially obscuring the impact of environmental gradients. These abnormalities appear in ordination but are in fact a product of oversaturated large pairwise distances.

RESULTS

We present Local Manifold distance (LMdist), an unsupervised algorithm which adjusts pairwise beta diversity measures to better represent true ecological distances between samples. Beta diversity measures can have a bounded dynamic range in depicting long environmental gradients with high species turnover. Using a graph structure, LMdist projects pairwise distances onto a manifold and traverses the manifold surface to adjust pairwise distances at the upper end of the beta diversity measure's dynamic range. This allows for values beyond the range of the original measure. Not all datasets will have oversaturated pairwise distances, nor will capture variation that resembles a manifold, so LMdist adjusts only those pairwise values which may be undervalued in the presence of a sampled gradient. The adjusted distances serve as input for ordination and statistical testing. We demonstrate on real and simulated data that LMdist effectively recovers distances along known gradients and along complex manifolds such as the Swiss roll dataset. LMdist enables more powerful statistical tests for gradient effects and reveals variation orthogonal to the gradient.

AVAILABILITY AND IMPLEMENTATION

Available on GitHub at https://github.com/knights-lab/LMdist.

Cultural transmission of move choice in chess

The study of cultural evolution benefits from detailed analysis of cultural transmission in specific human domains. Chess provides a platform for understanding the transmission of knowledge due to its active community of players, precise behaviours and long-term records of high-quality data. In this paper, we perform an analysis of chess in the context of cultural evolution, describing multiple cultural factors that affect move choice. We then build a population-level statistical model of move choice in chess, based on the Dirichlet-multinomial likelihood, to analyse cultural transmission over decades of recorded games played by leading players. For moves made in specific positions, we evaluate the relative effects of frequency-dependent bias, success bias and prestige bias on the dynamics of move frequencies. We observe that negative frequency-dependent bias plays a role in the dynamics of certain moves, and that other moves are compatible with transmission under prestige bias or success bias. These apparent biases may reflect recent changes, namely the introduction of computer chess engines and online tournament broadcasts. Our analysis of chess provides insights into broader questions concerning how social learning biases affect cultural evolution.

Evidence supportive of a bacterial component in the etiology for Alzheimer's disease and for a temporal-spatial development of a pathogenic microbiome in the brain

Background

Over the last few decades, a growing body of evidence has suggested a role for various infectious agents in Alzheimer's disease (AD) pathogenesis. Despite diverse pathogens (virus, bacteria, fungi) being detected in AD subjects' brains, research has focused on individual pathogens and only a few studies investigated the hypothesis of a bacterial brain microbiome. We profiled the bacterial communities present in non-demented controls and AD subjects' brains.

Results

We obtained postmortem samples from the brains of 32 individual subjects, comprising 16 AD and 16 control age-matched subjects with a total of 130 samples from the frontal and temporal lobes and the entorhinal cortex. We used full-length 16S rRNA gene amplification with Pacific Biosciences sequencing technology to identify bacteria. We detected bacteria in the brains of both cohorts with the principal bacteria comprising Cutibacterium acnes (formerly Propionibacterium acnes) and two species each of Acinetobacter and Comamonas genera. We used a hierarchical Bayesian method to detect differences in relative abundance among AD and control groups. Because of large abundance variances, we also employed a new analysis approach based on the Latent Dirichlet Allocation algorithm, used in computational linguistics. This allowed us to identify five sample classes, each revealing a different microbiota. Assuming that samples represented infections that began at different times, we ordered these classes in time, finding that the last class exclusively explained the existence or non-existence of AD.

Conclusions

The AD-related pathogenicity of the brain microbiome seems to be based on a complex polymicrobial dynamic. The time ordering revealed a rise and fall of the abundance of C. acnes with pathogenicity occurring for an off-peak abundance level in association with at least one other bacterium from a set of genera that included Methylobacterium, Bacillus, Caulobacter, Delftia, and Variovorax. C. acnes may also be involved with outcompeting the Comamonas species, which were strongly associated with non-demented brain microbiota, whose early destruction could be the first stage of disease. Our results are also consistent with a leaky blood-brain barrier or lymphatic network that allows bacteria, viruses, fungi, or other pathogens to enter the brain.

Topic modeling for multi-omic integration in the human gut microbiome and implications for Autism

While healthy gut microbiomes are critical to human health, pertinent microbial processes remain largely undefined, partially due to differential bias among profiling techniques. By simultaneously integrating multiple profiling methods, multi-omic analysis can define generalizable microbial processes, and is especially useful in understanding complex conditions such as Autism. Challenges with integrating heterogeneous data produced by multiple profiling methods can be overcome using Latent Dirichlet Allocation (LDA), a promising natural language processing technique that identifies topics in heterogeneous documents. In this study, we apply LDA to multi-omic microbial data (16S rRNA amplicon, shotgun metagenomic, shotgun metatranscriptomic, and untargeted metabolomic profiling) from the stool of 81 children with and without Autism. We identify topics, or microbial processes, that summarize complex phenomena occurring within gut microbial communities. We then subset stool samples by topic distribution, and identify metabolites, specifically neurotransmitter precursors and fatty acid derivatives, that differ significantly between children with and without Autism. We identify clusters of topics, deemed "cross-omic topics", which we hypothesize are representative of generalizable microbial processes observable regardless of profiling method. Interpreting topics, we find each represents a particular diet, and we heuristically label each cross-omic topic as: healthy/general function, age-associated function, transcriptional regulation, and opportunistic pathogenesis.

Comparative Analysis of Core Microbiome Assignments: Implications for Ecological Synthesis

The concept of a core microbiome has been broadly used to refer to the consistent presence of a set of taxa across multiple samples within a given habitat. The assignment of taxa to core microbiomes can be performed by several methods based on the abundance and occupancy (i.e., detection across samples) of individual taxa. These approaches have led to methodological inconsistencies, with direct implications for ecological interpretation. Here, we reviewed a set of methods most commonly used to infer core microbiomes in divergent systems. We applied these methods using large data sets and analyzed simulations to determine their accuracy in core microbiome assignments. Our results show that core taxa assignments vary significantly across methods and data set types, with occupancy-based methods most accurately defining true core membership. We also found the ability of these methods to accurately capture core assignments to be contingent on the distribution of taxon abundance and occupancy in the data set. Finally, we provide specific recommendations for further studies using core taxa assignments and discuss the need for unifying methodical approaches toward data processing to advance ecological synthesis. IMPORTANCE Different methods are commonly used to assign core microbiome membership, leading to methodological inconsistencies across studies. In this study, we review a set of the most commonly used core microbiome assignment methods and compare their core assignments using both simulated and empirical data. We report inconsistent classifications from commonly applied core microbiome assignment methods. Furthermore, we demonstrate the implication that variable core assignments may have on downstream ecological interpretations. Although we still lack a standardized approach to core taxa assignments, our study provides a direction to properly test core assignment methods and offers advances in model parameterization and method choice across distinct data types.

Spatiotemporal and ontogenetic variation, microbial selection, and predicted Bd-inhibitory function in the skin-associated microbiome of a Rocky Mountain amphibian

Host-associated microbiomes play important roles in host health and pathogen defense. In amphibians, the skin-associated microbiota can contribute to innate immunity with potential implications for disease management. Few studies have examined season-long temporal variation in the amphibian skin-associated microbiome, and the interactions between bacteria and fungi on amphibian skin remain poorly understood. We characterize season-long temporal variation in the skin-associated microbiome of the western tiger salamander (Ambystoma mavortium) for both bacteria and fungi between sites and across salamander life stages. Two hundred seven skin-associated microbiome samples were collected from salamanders at two Rocky Mountain lakes throughout the summer and fall of 2018, and 127 additional microbiome samples were collected from lake water and lake substrate. We used 16S rRNA and ITS amplicon sequencing with Bayesian Dirichlet-multinomial regression to estimate the relative abundances of bacterial and fungal taxa, test for differential abundance, examine microbial selection, and derive alpha diversity. We predicted the ability of bacterial communities to inhibit the amphibian chytrid fungus Batrachochytrium dendrobatidis (Bd), a cutaneous fungal pathogen, using stochastic character mapping and a database of Bd-inhibitory bacterial isolates. For both bacteria and fungi, we observed variation in community composition through time, between sites, and with salamander age and life stage. We further found that temporal trends in community composition were specific to each combination of salamander age, life stage, and lake. We found salamander skin to be selective for microbes, with many taxa disproportionately represented relative to the environment. Salamander skin appeared to select for predicted Bd-inhibitory bacteria, and we found a negative relationship between the relative abundances of predicted Bd-inhibitory bacteria and Bd. We hope these findings will assist in the conservation of amphibian species threatened by chytridiomycosis and other emerging diseases.

High-resolution population estimation using household survey data and building footprints

The national census is an essential data source to support decision-making in many areas of public interest. However, this data may become outdated during the intercensal period, which can stretch up to several decades. In this study, we develop a Bayesian hierarchical model leveraging recent household surveys and building footprints to produce up-to-date population estimates. We estimate population totals and age and sex breakdowns with associated uncertainty measures within grid cells of approximately 100 m in five provinces of the Democratic Republic of the Congo, a country where the last census was completed in 1984. The model exhibits a very good fit, with an R² value of 0.79 for out-of-sample predictions of population totals at the microcensus-cluster level and 1.00 for age and sex proportions at the province level. This work confirms the benefits of combining household surveys and building footprints for high-resolution population estimation in countries with outdated censuses.

A lack of up-to-date population figures may hamper effective decision-making. Here, the authors develop a Bayesian model to estimate population data at high resolution in five provinces of the Democratic Republic of the Congo.

Whole-Genome Duplication and Host Genotype Affect Rhizosphere Microbial Communities

The composition of microbial communities found in association with plants is influenced by host phenotype and genotype. However, the ways in which specific genetic architectures of host plants shape microbiomes are unknown. Genome duplication events are common in the evolutionary history of plants and influence many important plant traits, and thus, they may affect associated microbial communities. Using experimentally induced whole-genome duplication (WGD), we tested the effect of WGD on rhizosphere bacterial communities in Arabidopsis thaliana. We performed 16S rRNA amplicon sequencing to characterize differences between microbiomes associated with specific host genetic backgrounds (Columbia versus Landsberg) and ploidy levels (diploid versus tetraploid). We modeled relative abundances of bacterial taxa using a hierarchical Bayesian approach. We found that host genetic background and ploidy level affected rhizosphere community composition. We then tested to what extent microbiomes derived from a specific genetic background or ploidy level affected plant performance by inoculating sterile seedlings with microbial communities harvested from a prior generation. We found a negative effect of the tetraploid Columbia microbiome on growth of all four plant genetic backgrounds. These findings suggest an interplay between host genetic background and ploidy level and bacterial community assembly with potential ramifications for host fitness. Given the prevalence of ploidy-level variation in both wild and managed plant populations, the effects on microbiomes of this aspect of host genetic architecture could be a widespread driver of differences in plant microbiomes. IMPORTANCE Plants influence the composition of their associated microbial communities, yet the underlying host-associated genetic determinants are typically unknown. Genome duplication events are common in the evolutionary history of plants and affect many plant traits. Using Arabidopsis thaliana, we characterized how whole-genome duplication affected the composition of rhizosphere bacterial communities and how bacterial communities associated with two host plant genetic backgrounds and ploidy levels affected subsequent plant growth. We observed an interaction between ploidy level and genetic background that affected both bacterial community composition and function. This research reveals how genome duplication, a widespread genetic feature of both wild and crop plant species, influences bacterial assemblages and affects plant growth.

mRNA-1273 and BNT162b2 mRNA vaccines have reduced neutralizing activity against the SARS-CoV-2 Omicron variant

The BNT162b2 (Pfizer-BioNTech) and mRNA-1273 (Moderna) vaccines generate potent neutralizing antibodies against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). However, the global emergence of SARS-CoV-2 variants with mutations in the spike protein, the principal antigenic target of these vaccines, has raised concerns over the neutralizing activity of vaccine-induced antibody responses. The Omicron variant, which emerged in November 2021, consists of over 30 mutations within the spike protein. Here, we used an authentic live virus neutralization assay to examine the neutralizing activity of the SARS-CoV-2 Omicron variant against mRNA vaccine-induced antibody responses. Following the 2nd dose, we observed a 30-fold reduction in neutralizing activity against the omicron variant. Through six months after the 2nd dose, none of the sera from naïve vaccinated subjects showed neutralizing activity against the Omicron variant. In contrast, recovered vaccinated individuals showed a 22-fold reduction with more than half of the subjects retaining neutralizing antibody responses. Following a booster shot (3rd dose), we observed a 14-fold reduction in neutralizing activity against the omicron variant and over 90% of boosted subjects showed neutralizing activity against the omicron variant. These findings show that a 3rd dose is required to provide robust neutralizing antibody responses against the Omicron variant.

A suite of rare microbes interacts with a dominant, heritable, fungal endophyte to influence plant trait expression

Endophytes are microbes that live, for at least a portion of their life history, within plant tissues. Endophyte assemblages are often composed of a few abundant taxa and many infrequently observed, low-biomass taxa that are, in a word, rare. The ways in which most endophytes affect host phenotype are unknown; however, certain dominant endophytes can influence plants in ecologically meaningful ways-including by affecting growth and immune system functioning. In contrast, the effects of rare endophytes on their hosts have been unexplored, including how rare endophytes might interact with abundant endophytes to shape plant phenotype. Here, we manipulate both the suite of rare foliar endophytes (including both fungi and bacteria) and Alternaria fulva-a vertically transmitted and usually abundant fungus-within the fabaceous forb Astragalus lentiginosus. We report that rare, low-biomass endophytes affected host size and foliar %N, but only when the heritable fungal endophyte (A. fulva) was not present. A. fulva also reduced plant size and %N, but these deleterious effects on the host could be offset by a negative association we observed between this heritable fungus and a foliar pathogen. These results demonstrate how interactions among endophytic taxa determine the net effects on host plants and suggest that the myriad rare endophytes within plant leaves may be more than a collection of uninfluential, commensal organisms, but instead have meaningful ecological roles.

Host Immunity Alters Community Ecology and Stability of the Microbiome in a Caenorhabditis elegans Model

A growing body of data suggests that the microbiome of a species can vary considerably from individual to individual, but the reasons for this variation—and the consequences for the ecology of these communities—remain only partially explained. In mammals, the emerging picture is that the metabolic state and immune system status of the host affect the composition of the microbiome, but quantitative ecological microbiome studies are challenging to perform in higher organisms. Here, we show that these phenomena can be quantitatively analyzed in the tractable nematode host Caenorhabditis elegans. Mutants in innate immunity, in particular the DAF-2/insulin growth factor (IGF) pathway, are shown to contain a microbiome that differs from that of wild-type nematodes. We analyzed the underlying basis of these differences from the perspective of community ecology by comparing experimental observations to the predictions of a neutral sampling model and concluded that fundamental differences in microbiome ecology underlie the observed differences in microbiome composition. We tested this hypothesis by introducing a minor perturbation into the colonization conditions, allowing us to assess stability of communities in different host strains. Our results show that altering host immunity changes the importance of interspecies interactions within the microbiome, resulting in differences in community composition and stability that emerge from these differences in host-microbe ecology.

IMPORTANCE Here, we used a Caenorhabditis elegans microbiome model to demonstrate how genetic differences in innate immunity alter microbiome composition, diversity, and stability by changing the ecological processes that shape these communities. These results provide insight into the role of host genetics in controlling the ecology of the host-associated microbiota, resulting in differences in community composition, successional trajectories, and response to perturbation.

The quest for absolute abundance: The use of internal standards for DNA-based community ecology

To characterize microbiomes and other ecological assemblages, ecologists routinely sequence and compare loci that differ among focal taxa. Counts of these sequences convey information regarding the occurrence and relative abundances of taxa, but provide no direct measure of their absolute abundances, due to the technical limitations of the sequencing process. The relative abundances in compositional data are inherently constrained and difficult to interpret. The incorporation of internal standards (ISDs; colloquially referred to as 'spike-ins') into DNA pools can ameliorate the problems posed by relative abundance data and allow absolute abundances to be approximated. Unfortunately, many laboratory and sampling biases cause ISDs to underperform or fail. Here, we discuss how careful deployment of ISDs can avoid these complications and be an integral component of well-designed studies seeking to characterize ecological assemblages via sequencing of DNA.

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.

Applications of weighted association networks applied to compositional data in biology

Next-generation sequencing technologies have generated, and continue to produce, an increasingly large corpus of biological data. The data generated are inherently compositional as they convey only relative information dependent upon the capacity of the instrument, experimental design and technical bias. There is considerable information to be gained through network analysis by studying the interactions between components within a system. Network theory methods using compositional data are powerful approaches for quantifying relationships between biological components and their relevance to phenotype, environmental conditions or other external variables. However, many of the statistical assumptions used for network analysis are not designed for compositional data and can bias downstream results. In this mini-review, we illustrate the utility of network theory in biological systems and investigate modern techniques while introducing researchers to frameworks for implementation. We overview (1) compositional data analysis, (2) data transformations and (3) network theory along with insight on a battery of network types including static-, temporal-, sample-specific- and differential-networks. The intention of this mini-review is not to provide a comprehensive overview of network methods, rather to introduce microbiology researchers to (semi)-unsupervised data-driven approaches for inferring latent structures that may give insight into biological phenomena or abstract mechanics of complex systems.