Bayesian biclustering for microbial metagenomic sequencing data via multinomial matrix factorization

A Generalized Bayesian Stochastic Block Model for Microbiome Community Detection

Advances in next-generation sequencing technology have enabled the high-throughput profiling of metagenomes and accelerated microbiome studies. Recently, there has been a rise in quantitative studies that aim to decipher the microbiome co-occurrence network and its underlying community structure based on metagenomic sequence data. Uncovering the complex microbiome community structure is essential to understanding the role of the microbiome in disease progression and susceptibility. Taxonomic abundance data generated from metagenomic sequencing technologies are high-dimensional and compositional, suffering from uneven sampling depth, over-dispersion, and zero-inflation. These characteristics often challenge the reliability of the current methods for microbiome community detection. To study the microbiome co-occurrence network and perform community detection, we propose a generalized Bayesian stochastic block model that is tailored for microbiome data analysis where the data are transformed using the recently developed modified centered-log ratio transformation. Our model also allows us to leverage taxonomic tree information using a Markov random field prior. The model parameters are jointly inferred by using Markov chain Monte Carlo sampling techniques. Our simulation study showed that the proposed approach performs better than competing methods even when taxonomic tree information is non-informative. We applied our approach to a real urinary microbiome dataset from postmenopausal women. To the best of our knowledge, this is the first time the urinary microbiome co-occurrence network structure in postmenopausal women has been studied. In summary, this statistical methodology provides a new tool for facilitating advanced microbiome studies.

An Explainable Graph Neural Framework to Identify Cancer-Associated Intratumoral Microbial Communities

Microbes are extensively present among various cancer tissues and play critical roles in carcinogenesis and treatment responses. However, the underlying relationships between intratumoral microbes and tumors remain poorly understood. Here, a MIcrobial Cancer-association Analysis using a Heterogeneous graph transformer (MICAH) to identify intratumoral cancer-associated microbial communities is presented. MICAH integrates metabolic and phylogenetic relationships among microbes into a heterogeneous graph representation. It uses a graph transformer to holistically capture relationships between intratumoral microbes and cancer tissues, which improves the explainability of the associations between identified microbial communities and cancers. MICAH is applied to intratumoral bacterial data across 5 cancer types and 5 fungi datasets, and its generalizability and reproducibility are demonstrated. After experimentally testing a representative observation using a mouse model of tumor-microbe-immune interactions, a result consistent with MICAH's identified relationship is observed. Source tracking analysis reveals that the primary known contributor to a cancer-associated microbial community is the organs affected by the type of cancer. Overall, this graph neural network framework refines the number of microbes that can be used for follow-up experimental validation from thousands to tens, thereby helping to accelerate the understanding of the relationship between tumors and intratumoral microbiomes.

Analysis of Microbiome Data

The microbiome represents a hidden world of tiny organisms populating not only our surroundings but also our own bodies. By enabling comprehensive profiling of these invisible creatures, modern genomic sequencing tools have given us an unprecedented ability to characterize these populations and uncover their outsize impact on our environment and health. Statistical analysis of microbiome data is critical to infer patterns from the observed abundances. The application and development of analytical methods in this area require careful consideration of the unique aspects of microbiome profiles. We begin this review with a brief overview of microbiome data collection and processing and describe the resulting data structure. We then provide an overview of statistical methods for key tasks in microbiome data analysis, including data visualization, comparison of microbial abundance across groups, regression modeling, and network inference. We conclude with a discussion and highlight interesting future directions.

A Unified Bayesian Framework for Bi-overlapping-Clustering Multi-omics Data via Sparse Matrix Factorization

The advances of modern sequencing techniques have generated an unprecedented amount of multi-omics data which provide great opportunities to quantitatively explore functional genomes from different but complementary perspectives. However, distinct modalities/sequencing technologies generate diverse types of data which greatly complicate statistical modeling because uniquely optimized methods are required for handling each type of data. In this paper, we propose a unified framework for Bayesian nonparametric matrix factorization that infers overlapping bi-clusters for multi-omics data. The proposed method adaptively discretizes different types of observations into common latent states on which cluster structures are built hierarchically. The proposed Bayesian nonparametric method is able to automatically determine the number of clusters. We demonstrate the utility of the proposed method using simulation studies and applications to a single-cell RNA-sequencing dataset, a combination of single-cell RNA-sequencing and single-cell ATAC-sequencing dataset, a bulk RNA-sequencing dataset, and a DNA methylation dataset which reveal several interesting findings that are consistent with biological literature.

Lactobacillus fermentum CKCC1858 alleviates hyperlipidemia in golden hamsters on a high-fat diet via modulating gut microbiota

To investigate the effect of probiotic Lactobacillus fermentum CKCC1858, LF on the prevention of hyperlipidemia and its correlation with gut microbiota, golden hamsters were fed a high-fat diet alone or in combination with the probiotic for 6 weeks. The results showed that the LF intervention alleviated HFD-induced hyperlipidemia and liver damage, as evidenced by the reduced serum lipid profile levels and liver function markers. More importantly, the LF intervention attenuated HFD-induced microbiota dysbiosis by enhancing the abundance of SCFA-producing bacteria and reshaping the metabolic functions of the gut microbiota, likely contributing to its pronounced preventive effects on hyperlipidemia. This study elucidated the mechanism of the preventive effect of probiotics on hyperlipidemia in terms of regulating gut microbiota, and provided suggestions for regulating gut microbiota through probiotic interventions to improve lipid metabolism.

Inference of disease-associated microbial gene modules based on metagenomic and metatranscriptomic data

The identification of microbial characteristics associated with diseases is crucial for disease diagnosis and therapy. However, the presence of heterogeneity, high dimensionality, and large amounts of microbial data presents tremendous challenges in discovering key microbial features. In this paper, we present IDAM, a novel computational method for inferring disease-associated gene modules from metagenomic and metatranscriptomic data. This method integrates gene context conservation (uber-operons) and regulatory mechanisms (gene co-expression patterns) within a mathematical graph model to explore gene modules associated with specific diseases. It alleviates reliance on prior meta-data. We applied IDAM to publicly available datasets from inflammatory bowel disease, melanoma, type 1 diabetes mellitus, and irritable bowel syndrome. The results demonstrated the superior performance of IDAM in inferring disease-associated characteristics compared to existing popular tools. Furthermore, we showcased the high reproducibility of the gene modules inferred by IDAM using independent cohorts with inflammatory bowel disease. We believe that IDAM can be a highly advantageous method for exploring disease-associated microbial characteristics. The source code of IDAM is freely available at https://github.com/OSU-BMBL/IDAM, and the web server can be accessed at https://bmblx.bmi.osumc.edu/idam/.

Probiotics (Lactobacillus plantarum HNU082) Supplementation Relieves Ulcerative Colitis by Affecting Intestinal Barrier Functions, Immunity-Related Gene Expression, Gut Microbiota, and Metabolic Pathways in Mice

Probiotics can effectively improve ulcerative colitis (UC), but the mechanism is still unclear. Here, shotgun metagenome and transcriptome analyses were performed to explore the therapeutic effect and the mechanism of the probiotic Lactobacillus plantarum HNU082 (Lp082) on UC. The results showed that Lp082 treatment significantly ameliorated dextran sulfate sodium (DSS)-induced UC in mice, which was manifested as increases in body weight, water intake, food intake, and colon length and decreases in disease activity index (DAI), immune organ index, inflammatory factors, and histopathological scores after Lp082 intake. An in-depth study discovered that Lp082 could improve the intestinal mucosal barrier and relieve inflammation by cooptimizing the biological barrier, chemical barrier, mechanical barrier, and immune barrier. Specifically, Lp082 rebuilt the biological barrier by regulating the intestinal microbiome and increasing the production of short-chain fatty acids (SCFAs). Lp082 improved the chemical barrier by reducing intercellular cell adhesion molecule-1 (ICAM-1) and vascular cell adhesion molecule (VCAM) and increasing goblet cells and mucin2. Lp082 ameliorated the mechanical barrier by increasing zonula occludens-1 (ZO-1), zonula occludens-2 (ZO-2), and occludin while decreasing claudin-1 and claudin-2. Lp082 optimized the immune barrier by reducing the content of interleukin-1β (IL-1β), IL-6, tumor necrosis factor-α (TNF-α), myeloperoxidase (MPO), and interferon-γ (IFN-γ) and increasing IL-10, transforming growth factor-β1 (TGF-β1), and TGF-β2, inhibiting the NF-κB signaling pathway. Taken together, probiotic Lp082 can play a protective role in a DSS-induced colitis mouse model by protecting the intestinal mucosal barrier, attenuating the inflammatory response, and regulating microbial imbalance. This study provides support for the development of probiotic-based microbial products as an alternative treatment strategy for UC. IMPORTANCE Many studies have focused on the therapeutic effect of probiotics on ulcerative colitis (UC), but few studies have paid attention to the mechanism of probiotics, especially the therapeutic effect. This study suggests that Lp082 has a therapeutic effect on colitis in mice. Its mechanisms of action include protecting the mucosal barrier and actively modulating the gut microbiome, modulating inflammatory pathways, and reducing neutrophil infiltration. Our study enriches the mechanism and provides a new prospect for probiotics in the treatment of colitis, helps to deepen the understanding of the intestinal mucosal barrier, and provides guidance for the future probiotic treatment of human colitis.

Feature selection and classification over the network with missing node observations

Jointly analyzing transcriptomic data and the existing biological networks can yield more robust and informative feature selection results, as well as better understanding of the biological mechanisms. Selecting and classifying node features over genome-scale networks has become increasingly important in genomic biology and genomic medicine. Existing methods have some critical drawbacks. The first is they do not allow flexible modeling of different subtypes of selected nodes. The second is they ignore nodes with missing values, very likely to increase bias in estimation. To address these limitations, we propose a general modeling framework for Bayesian node classification (BNC) with missing values. A new prior model is developed for the class indicators incorporating the network structure. For posterior computation, we resort to the Swendsen-Wang algorithm for efficiently updating class indicators. BNC can naturally handle missing values in the Bayesian modeling framework, which improves the node classification accuracy and reduces the bias in estimating gene effects. We demonstrate the advantages of our methods via extensive simulation studies and the analysis of the cutaneous melanoma dataset from The Cancer Genome Atlas.

Host Genome-Metagenome Analyses Using Combinatorial Network Methods Reveal Key Metagenomic and Host Genetic Features for Methane Emission and Feed Efficiency in Cattle

Cattle production is one of the key contributors to global warming due to methane emission, which is a by-product of converting feed stuff into milk and meat for human consumption. Rumen hosts numerous microbial communities that are involved in the digestive process, leading to notable amounts of methane emission. The key factors underlying differences in methane emission between individual animals are due to, among other factors, both specific enrichments of certain microbial communities and host genetic factors that influence the microbial abundances. The detection of such factors involves various biostatistical and bioinformatics methods. In this study, our main objective was to reanalyze a publicly available data set using our proprietary Synomics Insights platform that is based on novel combinatorial network and machine learning methods to detect key metagenomic and host genetic features for methane emission and residual feed intake (RFI) in dairy cattle. The other objective was to compare the results with publicly available standard tools, such as those found in the microbiome bioinformatics platform QIIME2 and classic GWAS analysis. The data set used was publicly available and comprised 1,016 dairy cows with 16S short read sequencing data from two dairy cow breeds: Holstein and Nordic Reds. Host genomic data consisted of both 50 k and 150 k SNP arrays. Although several traits were analyzed by the original authors, here, we considered only methane emission as key phenotype for associating microbial communities and host genetic factors. The Synomics Insights platform is based on combinatorial methods that can identify taxa that are differentially abundant between animals showing high or low methane emission or RFI. Focusing exclusively on enriched taxa, for methane emission, the study identified 26 order-level taxa that combinatorial networks reported as significantly enriched either in high or low emitters. Additionally, a Z-test on proportions found 21/26 (81%) of these taxa were differentially enriched between high and low emitters (p value <.05). In particular, the phylum of Proteobacteria and the order Desulfovibrionales were found enriched in high emitters while the order Veillonellales was found to be more abundant in low emitters as previously reported for cattle (Wallace et al., 2015). In comparison, using the publicly available tool ANCOM only the order Methanosarcinales could be identified as differentially abundant between the two groups. We also investigated a link between host genome and rumen microbiome by applying our Synomics Insights platform and comparing it with an industry standard GWAS method. This resulted in the identification of genetic determinants in cows that are associated with changes in heritable components of the rumen microbiome. Only four key SNPs were found by both our platform and GWAS, whereas the Synomics Insights platform identified 1,290 significant SNPs that were not found by GWAS. Gene Ontology (GO) analysis found transcription factor as the dominant biological function. We estimated heritability of a core 73 taxa from the original set of 150 core order-level taxonomies and showed that some species are medium to highly heritable (0.25–0.62), paving the way for selective breeding of animals with desirable core microbiome characteristics. We identified a set of 113 key SNPs associated with >90% of these core heritable taxonomies. Finally, we have characterized a small set (<10) of SNPs strongly associated with key heritable bacterial orders with known role in methanogenesis, such as Desulfobacterales and Methanobacteriales.

Healthcare Biclustering-Based Prediction on Gene Expression Dataset

In this paper, we develop a healthcare biclustering model in the field of healthcare to reduce the inconveniences linked to the data clustering on gene expression. The present study uses two separate healthcare biclustering approaches to identify specific gene activity in certain environments and remove the duplication of broad gene information components. Moreover, because of its adequacy in the problem where populations of potential solutions allow exploration of a greater portion of the research area, machine learning or heuristic algorithm has become extensively used for healthcare biclustering in the field of healthcare. The study is evaluated in terms of average match score for nonoverlapping modules, overlapping modules through the influence of noise for constant bicluster and additive bicluster, and the run time. The results show that proposed FCM blustering method has higher average match score, and reduced run time proposed FCM than the existing PSO-SA and fuzzy logic healthcare biclustering methods.

Xylo-Oligosaccharides in Prevention of Hepatic Steatosis and Adipose Tissue Inflammation: Associating Taxonomic and Metabolomic Patterns in Fecal Microbiomes with Biclustering

We have shown that prebiotic xylo-oligosaccharides (XOS) increased beneficial gut microbiota (GM) and prevented high fat diet-induced hepatic steatosis, but the mechanisms associated with these effects are not clear. We studied whether XOS affects adipose tissue inflammation and insulin signaling, and whether the GM and fecal metabolome explain associated patterns. XOS was supplemented or not with high (HFD) or low (LFD) fat diet for 12 weeks in male Wistar rats (n = 10/group). Previously analyzed GM and fecal metabolites were biclustered to reduce data dimensionality and identify interpretable groups of co-occurring genera and metabolites. Based on our findings, biclustering provides a useful algorithmic method for capturing such joint signatures. On the HFD, XOS-supplemented rats showed lower number of adipose tissue crown-like structures, increased phosphorylation of AKT in liver and adipose tissue as well as lower expression of hepatic miRNAs. XOS-supplemented rats had more fecal glycine and less hypoxanthine, isovalerate, branched chain amino acids and aromatic amino acids. Several bacterial genera were associated with the metabolic signatures. In conclusion, the beneficial effects of XOS on hepatic steatosis involved decreased adipose tissue inflammation and likely improved insulin signaling, which were further associated with fecal metabolites and GM.

Applications of Machine Learning in Human Microbiome Studies: A Review on Feature Selection, Biomarker Identification, Disease Prediction and Treatment

The number of microbiome-related studies has notably increased the availability of data on human microbiome composition and function. These studies provide the essential material to deeply explore host-microbiome associations and their relation to the development and progression of various complex diseases. Improved data-analytical tools are needed to exploit all information from these biological datasets, taking into account the peculiarities of microbiome data, i.e., compositional, heterogeneous and sparse nature of these datasets. The possibility of predicting host-phenotypes based on taxonomy-informed feature selection to establish an association between microbiome and predict disease states is beneficial for personalized medicine. In this regard, machine learning (ML) provides new insights into the development of models that can be used to predict outputs, such as classification and prediction in microbiology, infer host phenotypes to predict diseases and use microbial communities to stratify patients by their characterization of state-specific microbial signatures. Here we review the state-of-the-art ML methods and respective software applied in human microbiome studies, performed as part of the COST Action ML4Microbiome activities. This scoping review focuses on the application of ML in microbiome studies related to association and clinical use for diagnostics, prognostics, and therapeutics. Although the data presented here is more related to the bacterial community, many algorithms could be applied in general, regardless of the feature type. This literature and software review covering this broad topic is aligned with the scoping review methodology. The manual identification of data sources has been complemented with: (1) automated publication search through digital libraries of the three major publishers using natural language processing (NLP) Toolkit, and (2) an automated identification of relevant software repositories on GitHub and ranking of the related research papers relying on learning to rank approach.