Variable selection in microbiome compositional data analysis

Microbiome compositional data analysis for survival studies

The growing interest in studying the relationship between the human microbiome and our health has also extended to time-to-event studies where researchers explore the connection between the microbiome and the occurrence of a specific event of interest. The analysis of microbiome obtained through high throughput sequencing techniques requires the use of specialized Compositional Data Analysis (CoDA) methods designed to accommodate its compositional nature. There is a limited availability of statistical tools for microbiome analysis that incorporate CoDA, and this is even more pronounced in the context of survival analysis. To fill this methodological gap, we present coda4microbiome for survival studies, a new methodology for the identification of microbial signatures in time-to-event studies. The algorithm implements an elastic-net penalized Cox regression model adapted to compositional covariates. We illustrate coda4microbiome algorithm for survival studies with a case study about the time to develop type 1 diabetes for non-obese diabetic mice. Our algorithm identified a bacterial signature composed of 21 genera associated with diabetes development. coda4microbiome for survival studies is integrated in the R package coda4microbiome as an extension of the existing functions for cross-sectional and longitudinal studies.

Gut microbiota disturbances in hospitalized older adults with malnutrition and clinical outcomes

OBJECTIVE

Malnutrition is one of the most threatening conditions in geriatric populations. The gut microbiota has an important role in the host's metabolic and muscular health: however, its interplay with disease-related malnutrition is not well understood. We aimed to identify the association of malnutrition with the gut microbiota and predict clinical outcomes in hospitalized acutely ill older adults.

METHODS

We performed a secondary longitudinal analysis in 108 geriatric patients from a prospective cohort evaluated at admission and 72 h of hospitalization. We collected clinical, demographic, nutritional, and 16S rRNA gene-sequenced gut microbiota data. Microbiota diversity, overall composition, and differential abundance were calculated and compared between patients with and without malnutrition. Microbiota features associated with malnutrition were used to predict clinical outcomes.

RESULTS

Patients with malnutrition (51%) had a different microbiota composition compared to those who were well-nourished during hospitalization (ANOSIM R = 0.079, P = 0.003). Patients with severe malnutrition showed poorer α-diversity at admission (Shannon P = 0.012, Simpson P = 0.018) and follow-up (Shannon P = 0.023, Chao1 P = 0.008). Differential abundance of Lachnospiraceae NK4A136 group, Subdoligranulum, and Faecalibacterium prausnitzii were significantly lower and inversely associated with malnutrition, while Corynebacterium, Ruminococcaceae Incertae Sedis, and Fusobacterium were significantly increased and positively associated with malnutrition. Corynebacterium, Ruminococcaceae Incertae Sedis, and the overall composition were important predictors of critical care in patients with malnutrition during hospitalization.

CONCLUSION

Older adults with malnutrition, especially in a severe stage, may be subject to substantial gut microbial disturbances during hospitalization. The gut microbiota profile of patients with malnutrition might help us to predict worse clinical outcomes.

Compositional structural brain signatures capture Alzheimer's genetic risk on brain structure along the disease continuum

INTRODUCTION

Traditional brain imaging genetics studies have primarily focused on how genetic factors influence the volume of specific brain regions, often neglecting the overall complexity of brain architecture and its genetic underpinnings.

METHODS

This study analyzed data from participants across the Alzheimer's disease (AD) continuum from the ALFA and ADNI studies. We exploited compositional data analysis to examine relative brain volumetric variations that (i) differentiate cognitively unimpaired (CU) individuals, defined as amyloid-negative (A-) based on CSF profiling, from those at different AD stages, and (ii) associated with increased genetic susceptibility to AD, assessed using polygenic risk scores.

RESULTS

Distinct brain signatures differentiated CU A-individuals from amyloid-positive MCI and AD. Moreover, disease stage-specific signatures were associated with higher genetic risk of AD.

DISCUSSION

The findings underscore the complex interplay between genetics and disease stages in shaping brain structure, which could inform targeted preventive strategies and interventions in preclinical AD.

Cotton microbiome profiling and Cotton Leaf Curl Disease (CLCuD) suppression through microbial consortia associated with Gossypium arboreum

The failure of breeding strategies has caused scientists to shift to other means where the new approach involves exploring the microbiome to modulate plant defense mechanisms against Cotton Leaf Curl Disease (CLCuD). The cotton microbiome of CLCuD-resistant varieties may harbor a multitude of bacterial genera that significantly contribute to disease resistance and provide information on metabolic pathways that differ between the susceptible and resistant varieties. The current study explores the microbiome of CLCuD-susceptible Gossypium hirsutum and CLCuD-resistant Gossypium arboreum using 16 S rRNA gene amplification for the leaf endophyte, leaf epiphyte, rhizosphere, and root endophyte of the two cotton species. This revealed that Pseudomonas inhabited the rhizosphere while Bacillus was predominantly found in the phyllosphere of CLCuV-resistant G. arboreum. Using salicylic acid-producing Serratia spp. and Fictibacillus spp. isolated from CLCuD-resistant G. arboreum, and guided by our analyses, we have successfully suppressed CLCuD in the susceptible G. hirsutum through pot assays. The applied strains exhibited less than 10% CLCuD incidence as compared to control group where it was 40% at 40 days post viral inoculation. Through detailed analytics, we have successfully demonstrated that the applied microbes serve as a biocontrol agent to suppress viral disease in Cotton.

Health-related quality of life is linked to the gut microbiome in kidney transplant recipients

Kidney transplant recipients (KTR) have impaired health-related quality of life (HRQoL) and suffer from intestinal dysbiosis. Increasing evidence shows that gut health and HRQoL are tightly related in the general population. Here, we investigate the association between the gut microbiome and HRQoL in KTR, using metagenomic sequencing data from fecal samples collected from 507 KTR. Multiple bacterial species are associated with lower HRQoL, many of which have previously been associated with adverse health conditions. Gut microbiome distance to the general population is highest among KTR with an impaired physical HRQoL (R = -0.20, P = 2.3 × 10-65) and mental HRQoL (R = -0.14, P = 1.3 × 10-3). Physical and mental HRQoL explain a significant part of variance in the gut microbiome (R2 = 0.58%, FDR = 5.43 × 10-4 and R2 = 0.37%, FDR = 1.38 × 10-3, respectively). Additionally, multiple metabolic and neuroactive pathways (gut brain modules) are associated with lower HRQoL. While the observational design of our study does not allow us to analyze causality, we provide a comprehensive overview of the associations between the gut microbiome and HRQoL while controlling for confounders.

A comprehensive analysis of gut and skin microbiota in canine atopic dermatitis in Shiba Inu dogs

BACKGROUND

Like its human counterpart, canine atopic dermatitis (cAD) is a chronic relapsing condition; thus, most cAD-affected dogs will require lifelong treatment to maintain an acceptable quality of life. A potential intervention is modulation of the composition of gut microbiota, and in fact, probiotic treatment has been proposed and tried in human atopic dermatitis (AD) patients. Since dogs are currently receiving intensive medical care, this will be the same option for dogs, while evidence of gut dysbiosis in cAD is still missing, although skin microbial profiling in cAD has been conducted in several studies. Therefore, we conducted a comprehensive analysis of both gut and skin microbiota in cAD in one specific cAD-predisposed breed, Shiba Inu. Additionally, we evaluated the impact of commonly used medical management on cAD (Janus kinase; JAK inhibitor, oclacitinib) on the gut and skin microbiota. Furthermore, we genotyped the Shiba Inu dogs according to the mitochondrial DNA haplogroup and assessed its association with the composition of the gut microbiota.

RESULTS

Staphylococcus was the most predominant bacterial genus observed in the skin; Escherichia/Shigella and Clostridium sensu stricto were highly abundant in the gut of cAD-affected dogs. In the gut microbiota, Fusobacteria and Megamonas were highly abundant in healthy dogs but significantly reduced in cAD-affected dogs. The abundance of these bacterial taxa was positively correlated with the effect of the treatment and state of the disease. Oclacitinib treatment on cAD-affected dogs shifted the composition of microbiota towards that in healthy dogs, and the latter brought it much closer to healthy microbiota, particularly in the gut. Additionally, even within the same dog breed, the mtDNA haplogroup varied, and there was an association between the mtDNA haplogroup and microbial composition in the gut and skin.

CONCLUSIONS

Dysbiosis of both the skin and the gut was observed in cAD in Shiba Inu dogs. Our findings provide a basis for the potential treatment of cAD by manipulating the gut microbiota as well as the skin microbiota. Video Abstract.

Effect of a Hop Extract Standardized in 8-Prenylnaringenin on Bone Health and Gut Microbiome in Postmenopausal Women with Osteopenia: A One-Year Randomized, Double-Blind, Placebo-Controlled Trial

Estrogen deficiency increases the risk of osteoporosis and fracture. The aim of this study was to investigate whether a hop extract standardized in 8-prenylnaringenin (8-PN), a potent phytoestrogen, could improve bone status of osteopenic women and to explore the gut microbiome roles in this effect. In this double-blind, placebo-controlled, randomized trial, 100 postmenopausal, osteopenic women were supplemented with calcium and vitamin D3 (CaD) tablets and either a hop extract (HE) standardized in 8-PN (n = 50) or a placebo (n = 50) for 48 weeks. Bone mineral density (BMD) and bone metabolism were assessed by DXA measurements and plasma bone biomarkers, respectively. Participant's quality of life (SF-36), gut microbiome composition, and short-chain fatty acid (SCFA) levels were also investigated. In addition to the CaD supplements, 48 weeks of HE supplementation increased total body BMD (1.8 ± 0.4% vs. baseline, p < 0.0001; 1.0 ± 0.6% vs. placebo, p = 0.08), with a higher proportion of women experiencing an increase ≥1% compared to placebo (odds ratio: 2.41 ± 1.07, p < 0.05). An increase in the SF-36 physical functioning score was observed with HE versus placebo (p = 0.05). Gut microbiome α-diversity and SCFA levels did not differ between groups. However, a higher abundance of genera Turicibacter and Shigella was observed in the HE group; both genera have been previously identified as associated with total body BMD. These results suggest that an 8-PN standardized hop extract could beneficially impact bone health of postmenopausal women with osteopenia.

Molecular insights informing factors affecting low temperature anaerobic applications: Diversity, collated core microbiomes and complexity stability relationships in LCFA-fed systems

Fats, oil and grease, and their hydrolyzed counterparts-long chain fatty acids (LCFA) make up a large fraction of numerous wastewaters and are challenging to degrade anaerobically, more so, in low temperature anaerobic digestion (LtAD) systems. Herein, we perform a comparative analysis of publicly available Illumina 16S rRNA datasets generated from LCFA-degrading anaerobic microbiomes at low temperatures (10 and 20 °C) to comprehend the factors affecting microbial community dynamics. The various factors considered were the inoculum, substrate and operational characteristics, the reactor operation mode and reactor configuration, and the type of nucleic acid sequenced. We found that LCFA-degrading anaerobic microbiomes were differentiated primarily by inoculum characteristics (inoculum source and morphology) in comparison to the other factors tested. Inoculum characteristics prominently shaped the species richness, species evenness and beta-diversity patterns in the microbiomes even after long term operation of continuous reactors up to 150 days, implying the choice of inoculum needs careful consideration. The generalised additive models represented through beta diversity contour plots revealed that psychrophilic bacteria RBG-13-54-9 from family Anaerolineae, and taxa WCHB1-41 and Williamwhitmania were highly abundant in LCFA-fed microbial niches, suggesting their role in anaerobic treatment of LCFAs at low temperatures of 10-20 °C. Overall, we showed that the following bacterial genera: uncultured Propionibacteriaceae, Longilinea, Christensenellaceae R7 group, Lactivibrio, candidatus Caldatribacterium, Aminicenantales, Syntrophus, Syntrophomonas, Smithella, RBG-13-54-9, WCHB1-41, Trichococcus, Proteiniclasticum, SBR1031, Lutibacter and Lentimicrobium have prominent roles in LtAD of LCFA-rich wastewaters at 10-20 °C. This study provides molecular insights of anaerobic LCFA degradation under low temperatures from collated datasets and will aid in improving LtAD systems for treating LCFA-rich wastewaters.

PLSDA-batch: a multivariate framework to correct for batch effects in microbiome data

Microbial communities are highly dynamic and sensitive to changes in the environment. Thus, microbiome data are highly susceptible to batch effects, defined as sources of unwanted variation that are not related to and obscure any factors of interest. Existing batch effect correction methods have been primarily developed for gene expression data. As such, they do not consider the inherent characteristics of microbiome data, including zero inflation, overdispersion and correlation between variables. We introduce new multivariate and non-parametric batch effect correction methods based on Partial Least Squares Discriminant Analysis (PLSDA). PLSDA-batch first estimates treatment and batch variation with latent components, then subtracts batch-associated components from the data. The resulting batch-effect-corrected data can then be input in any downstream statistical analysis. Two variants are proposed to handle unbalanced batch x treatment designs and to avoid overfitting when estimating the components via variable selection. We compare our approaches with popular methods managing batch effects, namely, removeBatchEffect, ComBat and Surrogate Variable Analysis, in simulated and three case studies using various visual and numerical assessments. We show that our three methods lead to competitive performance in removing batch variation while preserving treatment variation, especially for unbalanced batch $\times $ treatment designs. Our downstream analyses show selections of biologically relevant taxa. This work demonstrates that batch effect correction methods can improve microbiome research outputs. Reproducible code and vignettes are available on GitHub.

coda4microbiome: compositional data analysis for microbiome cross-sectional and longitudinal studies

BACKGROUND

One of the main challenges of microbiome analysis is its compositional nature that if ignored can lead to spurious results. Addressing the compositional structure of microbiome data is particularly critical in longitudinal studies where abundances measured at different times can correspond to different sub-compositions.

RESULTS

We developed coda4microbiome, a new R package for analyzing microbiome data within the Compositional Data Analysis (CoDA) framework in both, cross-sectional and longitudinal studies. The aim of coda4microbiome is prediction, more specifically, the method is designed to identify a model (microbial signature) containing the minimum number of features with the maximum predictive power. The algorithm relies on the analysis of log-ratios between pairs of components and variable selection is addressed through penalized regression on the "all-pairs log-ratio model", the model containing all possible pairwise log-ratios. For longitudinal data, the algorithm infers dynamic microbial signatures by performing penalized regression over the summary of the log-ratio trajectories (the area under these trajectories). In both, cross-sectional and longitudinal studies, the inferred microbial signature is expressed as the (weighted) balance between two groups of taxa, those that contribute positively to the microbial signature and those that contribute negatively. The package provides several graphical representations that facilitate the interpretation of the analysis and the identified microbial signatures. We illustrate the new method with data from a Crohn's disease study (cross-sectional data) and on the developing microbiome of infants (longitudinal data).

CONCLUSIONS

coda4microbiome is a new algorithm for identification of microbial signatures in both, cross-sectional and longitudinal studies. The algorithm is implemented as an R package that is available at CRAN ( https://cran.r-project.org/web/packages/coda4microbiome/ ) and is accompanied with a vignette with a detailed description of the functions. The website of the project contains several tutorials: https://malucalle.github.io/coda4microbiome/.

Small Spatial Scale Drivers of Secondary Metabolite Biosynthetic Diversity in Environmental Microbiomes

In the search for novel drug candidates, diverse environmental microbiomes have been surveyed for their secondary metabolite biosynthesis potential, yet little is known about the biosynthetic diversity encoded by divergent microbiomes from different ecosystems, and the environmental parameters driving this diversity. Here, we used targeted amplicon sequencing of adenylation (AD) and ketosynthase (KS) domains along with 16S sequencing to delineate the unique biosynthetic potential of microbiomes from three separate habitats (soil, water, and sediments) exhibiting unique small spatial scale physicochemical gradients. The estimated richness of AD domains was highest in marine sediments with 656 ± 58 operational biosynthetic units (OBUs), while the KS domain richness was highest in soil microbiomes with 388 ± 67 OBUs. Microbiomes with rich and diverse bacterial communities displayed the highest PK potential across all ecosystems, and on a small spatial scale, pH and salinity were significantly, positively correlated to KS domain richness in soil and aquatic systems, respectively. Integrating our findings, we were able to predict the KS domain richness with a RMSE of 31 OBUs and a R² of 0.91, and by the use of publicly available information on bacterial richness and diversity, we identified grassland biomes as being particularly promising sites for the discovery of novel polyketides. Furthermore, a focus on acidobacterial taxa is likely to be fruitful, as these were responsible for most of the variation in biosynthetic diversity. Overall, our results highlight the importance of sampling diverse environments with high taxonomic diversity in the pursuit for novel secondary metabolites. IMPORTANCE To counteract the antibiotic resistance crisis, novel anti-infective agents need to be discovered and brought to market. Microbial secondary metabolites have been important sources of inspiration for small-molecule therapeutics. However, the isolation of novel antibiotics is difficult, and the risk of rediscovery is high. With the overarching purpose of identifying promising microbiomes for discovery of novel bioactivity, we mapped out the most significant drivers of biosynthetic diversity across divergent microbiomes. We found the biosynthetic potential to be unique to individual ecosystems, and to depend on bacterial taxonomic diversity. Within systems, and on small spatial scales, pH and salinity correlated positively to the biosynthetic richness of the microbiomes, Acidobacteria representing the taxa most highly associated with biosynthetic diversity. Ultimately, understanding the key drivers of the biosynthesis potential of environmental microbiomes will allow us to focus bioprospecting efforts and facilitate the discovery of novel therapeutics.

Resampling-based inferences for compositional regression with application to beef cattle microbiomes

Gut microbiomes are increasingly found to be associated with many health-related characteristics of humans as well as animals. Regression with compositional microbiomes covariates is commonly used to identify important bacterial taxa that are related to various phenotype responses. Often the dimension of microbiome taxa easily exceeds the number of available samples, which creates a serious challenge in the estimation and inference of the model. The sparse log-contrast regression method is useful for such cases as it can yield a model estimate that depends on only a small number of taxa. However, a formal statistical inference procedure for individual regression coefficients has not been properly established yet. We propose a new estimation and inference procedure for linear regression models with extremely low-sample-sized compositional predictors. Under the compositional log-contrast regression framework, the proposed approach consists of two steps. The first step is to screen relevant predictors by fitting a log-contrast model with a sparse penalty. The screened-in variables are used as predictors in the non-sparse log-contrast model in the second step, where each of the regression coefficients is tested using nonparametric, resampling-based methods such as permutation and bootstrap. The performances of the proposed methods are evaluated by a simulation study, which shows they outperform traditional approaches based on normal assumptions or large sample asymptotics. Application to steer microbiome data successfully identifies key bacterial taxa that are related to important cattle quality measures.

The diagnostic potential and barriers of microbiome based therapeutics

High throughput technological innovations in the past decade have accelerated research into the trillions of commensal microbes in the gut. The 'omics' technologies used for microbiome analysis are constantly evolving, and large-scale datasets are being produced. Despite of the fact that much of the research is still in its early stages, specific microbial signatures have been associated with the promotion of cancer, as well as other diseases such as inflammatory bowel disease, neurogenerative diareses etc. It has been also reported that the diversity of the gut microbiome influences the safety and efficacy of medicines. The availability and declining sequencing costs has rendered the employment of RNA-based diagnostics more common in the microbiome field necessitating improved data-analytical techniques so as to fully exploit all the resulting rich biological datasets, while accounting for their unique characteristics, such as their compositional nature as well their heterogeneity and sparsity. As a result, the gut microbiome is increasingly being demonstrating as an important component of personalised medicine since it not only plays a role in inter-individual variability in health and disease, but it also represents a potentially modifiable entity or feature that may be addressed by treatments in a personalised way. In this context, machine learning and artificial intelligence-based methods may be able to unveil new insights into biomedical analyses through the generation of models that may be used to predict category labels, and continuous values. Furthermore, diagnostic aspects will add value in the identification of the non invasive markers in the critical diseases like cancer.

Beyond Basic Diversity Estimates-Analytical Tools for Mechanistic Interpretations of Amplicon Sequencing Data

Understanding microbial ecology through amplifying short read regions, typically 16S rRNA for prokaryotic species or 18S rRNA for eukaryotic species, remains a popular, economical choice. These methods provide relative abundances of key microbial taxa, which, depending on the experimental design, can be used to infer mechanistic ecological underpinnings. In this review, we discuss recent advancements in in situ analytical tools that have the power to elucidate ecological phenomena, unveil the metabolic potential of microbial communities, identify complex multidimensional interactions between species, and compare stability and complexity under different conditions. Additionally, we highlight methods that incorporate various modalities and additional information, which in combination with abundance data, can help us understand how microbial communities respond to change in a typical ecosystem. Whilst the field of microbial informatics continues to progress substantially, our emphasis is on popular methods that are applicable to a broad range of study designs. The application of these methods can increase our mechanistic understanding of the ongoing dynamics of complex microbial communities.

Principal microbial groups: compositional alternative to phylogenetic grouping of microbiome data

Statistical and machine learning techniques based on relative abundances have been used to predict health conditions and to identify microbial biomarkers. However, high dimensionality, sparsity and the compositional nature of microbiome data represent statistical challenges. On the other hand, the taxon grouping allows summarizing microbiome abundance with a coarser resolution in a lower dimension, but it presents new challenges when correlating taxa with a disease. In this work, we present a novel approach that groups Operational Taxonomical Units (OTUs) based only on relative abundances as an alternative to taxon grouping. The proposed procedure acknowledges the compositional data making use of principal balances. The identified groups are called Principal Microbial Groups (PMGs). The procedure reduces the need for user-defined aggregation of $\textrm{OTU}$s and offers the possibility of working with coarse group of $\textrm{OTU}$s, which are not present in a phylogenetic tree. PMGs can be used for two different goals: (1) as a dimensionality reduction method for compositional data, (2) as an aggregation procedure that provides an alternative to taxon grouping for construction of microbial balances afterward used for disease prediction. We illustrate the procedure with a cirrhosis study data. PMGs provide a coherent data analysis for the search of biomarkers in human microbiota. The source code and demo data for PMGs are available at: https://github.com/asliboyraz/PMGs.

Three approaches to supervised learning for compositional data with pairwise logratios

Logratios between pairs of compositional parts (pairwise logratios) are the easiest to interpret in compositional data analysis, and include the well-known additive logratios as particular cases. When the number of parts is large (sometimes even larger than the number of cases), some form of logratio selection is needed. In this article, we present three alternative stepwise supervised learning methods to select the pairwise logratios that best explain a dependent variable in a generalized linear model, each geared for a specific problem. The first method features unrestricted search, where any pairwise logratio can be selected. This method has a complex interpretation if some pairs of parts in the logratios overlap, but it leads to the most accurate predictions. The second method restricts parts to occur only once, which makes the corresponding logratios intuitively interpretable. The third method uses additive logratios, so that K-1 selected logratios involve a K-part subcomposition. Our approach allows logratios or non-compositional covariates to be forced into the models based on theoretical knowledge, and various stopping criteria are available based on information measures or statistical significance with the Bonferroni correction. We present an application on a dataset from a study predicting Crohn's disease.

Predicting cancer immunotherapy response from gut microbiomes using machine learning models

Cancer immunotherapy has significantly improved patient survival. Yet, half of patients do not respond to immunotherapy. Gut microbiomes have been linked to clinical responsiveness of melanoma patients on immunotherapies; however, different taxa have been associated with response status with implicated taxa inconsistent between studies. We used a tumor-agnostic approach to find common gut microbiome features of response among immunotherapy patients with different advanced stage cancers. A combined meta-analysis of 16S rRNA gene sequencing data from our mixed tumor cohort and three published immunotherapy gut microbiome datasets from different melanoma patient cohorts found certain gut bacterial taxa correlated with immunotherapy response status regardless of tumor type. Using multivariate selbal analysis, we identified two separate groups of bacterial genera associated with responders versus non-responders. Statistical models of gut microbiome community features showed robust prediction accuracy of immunotherapy response in amplicon sequencing datasets and in cross-sequencing platform validation with shotgun metagenomic datasets. Results suggest baseline gut microbiome features may be predictive of clinical outcomes in oncology patients on immunotherapies, and some of these features may be generalizable across different tumor types, patient cohorts, and sequencing platforms. Findings demonstrate how machine learning models can reveal microbiome-immunotherapy interactions that may ultimately improve cancer patient outcomes.

Principal Amalgamation Analysis for Microbiome Data

In recent years microbiome studies have become increasingly prevalent and large-scale. Through high-throughput sequencing technologies and well-established analytical pipelines, relative abundance data of operational taxonomic units and their associated taxonomic structures are routinely produced. Since such data can be extremely sparse and high dimensional, there is often a genuine need for dimension reduction to facilitate data visualization and downstream statistical analysis. We propose Principal Amalgamation Analysis (PAA), a novel amalgamation-based and taxonomy-guided dimension reduction paradigm for microbiome data. Our approach aims to aggregate the compositions into a smaller number of principal compositions, guided by the available taxonomic structure, by minimizing a properly measured loss of information. The choice of the loss function is flexible and can be based on familiar diversity indices for preserving either within-sample or between-sample diversity in the data. To enable scalable computation, we develop a hierarchical PAA algorithm to trace the entire trajectory of successive simple amalgamations. Visualization tools including dendrogram, scree plot, and ordination plot are developed. The effectiveness of PAA is demonstrated using gut microbiome data from a preterm infant study and an HIV infection study.

A potential oral microbiome signature associated with coronary artery disease in Tunisia

The coronary artery disease is a chronic inflammatory disease involving genetic as well as environmental factors. Recent evidence suggests that the oral microbiome has a significant role in triggering atherosclerosis. The present study assessed the oral microbiome composition variation between coronary patients and healthy subjects in order to identify a potential pathogenic signature associated with coronary artery disease (CAD). We performed metagenomic profiling of salivary microbiomes by 16S rRNA next-generation sequencing. Oral microbiota profiling was performed for 30 individuals including 20 patients with CAD and 10 healthy individuals without carotid plaques or previous stroke or myocardial infarction.We found that oral microbial communities in patients and healthy controls are represented by similar global core oral microbiome. The predominant taxa belonged to Firmicutes (genus Streptococcus, Veillonella, Granulicatella, Selenomonas), Proteobacteria (genus Neisseria, Haemophilus), Actinobacteria (genus Rothia), Bacteroidetes (genus Prevotella, Porphyromonas) and Fusobacteria (genus Fusobacterium, Leptotrichia). More than 60% relative abundance of each sample for both CAD patients and controls is represented by three major genera including Streptococcus (24.97% and 26.33%), Veillonella (21.43% and 19.91%) and Neisseria (14.23% and 15.33%).  Using penalized regression analysis, the bacterial genus Eikenella was involved as the major discriminant genus for both status and Syntax score of CAD. We also reported a significant negative correlation between Syntax score and Eikenella abundance in coronary patients' group (Spearman rho =-0.68, p= 0.00094). In conclusion, the abundance of Eikenella in oral coronary patient samples compared to controls could be a prominent pathological indicator for the development of CAD.

A Simultaneous Feature Selection and Compositional Association Test for Detecting Sparse Associations in High-Dimensional Metagenomic Data

Numerous metagenomic studies aim to discover associations between the microbial composition of an environment (e.g., gut, skin, oral) and a phenotype of interest. Multivariate analysis is often performed in these studies without critical a priori knowledge of which taxa are associated with the phenotype being studied. This approach typically reduces statistical power in settings where the true associations among only a few taxa are obscured by high dimensionality (i.e., sparse association signals). At the same time, low sample size and compositional sample space constraints may reduce beyond-study generalizability if not properly accounted for. To address these difficulties, we developed the Selection-Energy-Permutation (SelEnergyPerm) method, a nonparametric group association test with embedded feature selection that directly accounts for compositional constraints using parsimonious logratio signatures between taxonomic features, for characterizing and understanding alterations in microbial community structure. Simulation results show SelEnergyPerm selects small independent sets of logratios that capture strong associations in a range of scenarios. Additionally, our simulation results demonstrate SelEnergyPerm consistently detects/rejects associations in synthetic data with sparse, dense, or no association signals. We demonstrate the novel benefits of our method in four case studies utilizing publicly available 16S amplicon and whole-genome sequencing datasets. Our R implementation of Selection-Energy-Permutation, including an example demonstration and the code to generate all of the scenarios used here, is available at https://www.github.com/andrew84830813/selEnergyPermR.

Adaptive and powerful microbiome multivariate association analysis via feature selection

The important role of human microbiome is being increasingly recognized in health and disease conditions. Since microbiome data is typically high dimensional, one popular mode of statistical association analysis for microbiome data is to pool individual microbial features into a group, and then conduct group-based multivariate association analysis. A corresponding challenge within this approach is to achieve adequate power to detect an association signal between a group of microbial features and the outcome of interest across a wide range of scenarios. Recognizing some existing methods' susceptibility to the adverse effects of noise accumulation, we introduce the Adaptive Microbiome Association Test (AMAT), a novel and powerful tool for multivariate microbiome association analysis, which unifies both blessings of feature selection in high-dimensional inference and robustness of adaptive statistical association testing. AMAT first alleviates the burden of noise accumulation via distance correlation learning, and then conducts a data-adaptive association test under the flexible generalized linear model framework. Extensive simulation studies and real data applications demonstrate that AMAT is highly robust and often more powerful than several existing methods, while preserving the correct type I error rate. A free implementation of AMAT in R computing environment is available at https://github.com/kzb193/AMAT.

Novel application of survival models for predicting microbial community transitions with variable selection for eDNA

Survival analysis is a prolific statistical tool in medicine for inferring risk and time to disease-related events. However, it is under-utilized in microbiome research to predict microbial community mediated events, partly due to the sparsity and high dimensional nature of the data. We advance the application of Cox proportional hazards (Cox PH) survival models to environmental DNA (eDNA) data with feature selection suitable for filtering irrelevant and redundant taxonomic variables. Selection methods are compared in terms of false positives, sensitivity, and survival estimation accuracy in simulation and in a real data setting to forecast harmful cyanobacterial blooms. A novel extension of a method for selecting microbial biomarkers with survival data (SuRFCox) reliably outperforms other methods. We determine Cox PH models with SuRFCox selected predictors are more robust to varied signal, noise, and data correlation structure. SuRFCox also yields the most accurate and consistent prediction of blooms according to cross-validated testing by year over eight different bloom seasons. Identification of common biomarkers among validated survival forecasts over changing conditions has clear biological significance. Survival models with such biomarkers inform risk assessment and provide insight into the causes of critical community transitions. Importance In this paper, we report on a novel approach of selecting microorganisms for model-based prediction of the time to critical microbially-modulated events (e.g., harmful algal blooms, clinical outcomes, community shifts, etc.). Our novel method for identifying biomarkers from large, dynamic communities of microbes has broad utility to environmental and ecological impact risk assessment and public health. Results will also promote theoretical and practical advancements relevant to the biology of specific organisms. To address the unique challenge posed by diverse environmental conditions and sparse microbes, we developed a novel method of selecting predictors for modelling time-to-event data. Competing methods for selecting predictors are rigorously compared to determine which is the most accurate and generalizable. Model forecasts are applied to show suitable predictors can precisely quantify the risk over time of biological events like harmful cyanobacterial blooms.

Evaluating supervised and unsupervised background noise correction in human gut microbiome data

The ability to predict human phenotypes and identify biomarkers of disease from metagenomic data is crucial for the development of therapeutics for microbiome-associated diseases. However, metagenomic data is commonly affected by technical variables unrelated to the phenotype of interest, such as sequencing protocol, which can make it difficult to predict phenotype and find biomarkers of disease. Supervised methods to correct for background noise, originally designed for gene expression and RNA-seq data, are commonly applied to microbiome data but may be limited because they cannot account for unmeasured sources of variation. Unsupervised approaches address this issue, but current methods are limited because they are ill-equipped to deal with the unique aspects of microbiome data, which is compositional, highly skewed, and sparse. We perform a comparative analysis of the ability of different denoising transformations in combination with supervised correction methods as well as an unsupervised principal component correction approach that is presently used in other domains but has not been applied to microbiome data to date. We find that the unsupervised principal component correction approach has comparable ability in reducing false discovery of biomarkers as the supervised approaches, with the added benefit of not needing to know the sources of variation apriori. However, in prediction tasks, it appears to only improve prediction when technical variables contribute to the majority of variance in the data. As new and larger metagenomic datasets become increasingly available, background noise correction will become essential for generating reproducible microbiome analyses.

The human gut microbiome is known to play a major role in health and is associated with many diseases including colorectal cancer, obesity, and diabetes. The prediction of host phenotypes and identification of biomarkers of disease is essential for harnessing the therapeutic potential of the microbiome. However, many metagenomic datasets are affected by technical variables that introduce unwanted variation that can confound the ability to predict phenotypes and identify biomarkers. Currently, supervised methods originally designed for gene expression and RNA-seq data are commonly applied to microbiome data for correction of background noise, but they are limited in that they cannot correct for unmeasured sources of variation. Unsupervised approaches address this issue, but current methods are limited because they are ill-equipped to deal with the unique aspects of microbiome data, which is compositional, highly skewed, and sparse. We perform a comparative analysis of the ability of different denoising transformations in combination with supervised correction methods as well as an unsupervised principal component correction approach and find that all correction approaches reduce false positives for biomarker discovery. In the task of predicting phenotypes, different approaches have varying success where the unsupervised correction can improve prediction when technical variables contribute to the majority of variance in the data. As new and larger metagenomic datasets become increasingly available, background noise correction will become essential for generating reproducible microbiome analyses.

Changes in Serum N-Glycome for Risk Drinkers: A Comparison with Standard Markers for Alcohol Abuse in Men and Women

Background and aim: Glycomic alterations serve as biomarker tools for different diseases. The present study aims to evaluate the diagnostic capability of serum N-glycosylation to identify alcohol risk drinking in comparison with standard markers. Methods: We included 1516 adult individuals (age range 18–91 years; 55.3% women), randomly selected from a general population. A total of 143 (21.0%) men and 50 (5.9%) women were classified as risk drinkers after quantification of daily alcohol consumption and the Alcohol Use Disorders Identification Test (AUDIT). Hydrophilic interaction ultra-performance liquid chromatography (HILIC-UPLC) was used for the quantification of 46 serum N-glycan peaks. Serum gamma-glutamyltransferase (GGT), carbohydrate-deficient transferrin (CDT), and red blood cell mean corpuscular volume (MCV) were measured by standard clinical laboratory methods. Results: Variations in serum N-glycome associated risk drinking were more prominent in men compared to women. A unique combination of N-glycan peaks selected by the selbal algorithm shows good discrimination between risk-drinkers and non-risk drinkers for men and women. Receiver operating characteristics (ROC) curves show accuracy for the diagnosis of risk drinking, which is comparable to that of the golden standards, GGT, MCV and CDT markers for men and women. Additionally, the inclusion of N-glycan peaks improves the diagnostic accuracy of the standard markers, although it remains relatively low, due to low sensitivity. For men, the area under the ROC curve using N-glycome data is 0.75, 0.76, and 0.77 when combined with GGT, MCV, and CDT, respectively. In women, the areas were 0.76, 0.73, and 0.73, respectively. Conclusion: Risk drinking is associated with significant variations in the serum N-glycome, which highlights its potential diagnostic utility.

Microbiome Analysis of Mucosal Ileoanal Pouch in Ulcerative Colitis Patients Revealed Impairment of the Pouches Immunometabolites

The pathogenesis of ulcerative colitis (UC) is unknown, although genetic loci and altered gut microbiota have been implicated. Up to a third of patients with moderate to severe UC require proctocolectomy with ileal pouch ano-anastomosis (IPAA). We aimed to explore the mucosal microbiota of UC patients who underwent IPAA.

METHODS

For microbiome analysis, mucosal specimens were collected from 34 IPAA individuals. Endoscopic and histological examinations of IPAA were normal in 21 cases, while pouchitis was in 13 patients. 19 specimens from the healthy control (10 from colonic and 9 from ileum) were also analyzed. Data were analyzed using an ensemble of software packages: QIIME2, coda-lasso, clr-lasso, PICRUSt2, and ALDEx2.

RESULTS

IPAA specimens had significantly lower bacterial diversity as compared to normal. The microbial composition of the normal pouch was also decreased also when compared to pouchitis. Faecalibacterium prausnitzii, Gemmiger formicilis, Blautia obeum, Ruminococcus torques, Dorea formicigenerans, and an unknown species from Roseburia were the most uncommon in pouch/pouchitis, while an unknown species from Enterobacteriaceae was over-represented. Propionibacterium acnes and Enterobacteriaceae were the species most abundant in the pouchitis and in the normal pouch, respectively. Predicted metabolic pathways among the IPAA bacterial communities revealed an important role of immunometabolites such as SCFA, butyrate, and amino acids.

CONCLUSIONS

Our findings showed specific bacterial signature hallmarks of dysbiosis and could represent bacterial biomarkers in IPAA patients useful to develop novel treatments in the future by modulating the gut microbiota through the administration of probiotic immunometabolites-producing bacterial strains and the addition of specific prebiotics and the faecal microbiota transplantation.

Multivariate statistical monitoring system for microbial population dynamics

Microbiomes in their natural environments vary dynamically with changing environmental conditions. The detection of these dynamic changes in microbial populations is critical for understanding the impact of environmental changes on the microbial community. Here, we propose a novel method to detect time-series changes in the microbiome, based on multivariate statistical process control. By focusing on the interspecies structures, this approach enables the robust detection of time-series changes in a microbiome composed of a large number of microbial species. Applying this approach to empirical human gut microbiome data, we accurately traced time-series changes in microbiota composition induced by a dietary intervention trial. This method was also excellent for tracking the recovery process after the intervention. Our approach can be useful for monitoring dynamic changes in complex microbial communities.

A comprehensive evaluation of single-end sequencing data analyses for environmental microbiome research

Illumina sequencing platforms have been widely used for amplicon-based environmental microbiome research. Analyses of amplicon data of environmental samples, generated from Illumina MiSeq platform illustrate the reverse (R2) reads in the PE datasets to have low quality towards the 3' end of the reads which affect the sequencing depth of samples and ultimately impact the sample size which may possibly lead to an altered outcome. This study evaluates the usefulness of single-end (SE) sequencing data in microbiome research when the Illumina MiSeq PE dataset shows significantly high number of low-quality reverse reads. In this study, the amplicon data (V1V3, V3V4, V4V5 and V6V8) from 128 environmental (soil) samples, downloaded from SRA, demonstrate the efficiency of single-end (SE) sequencing data analyses in microbiome research. The SE datasets were found to infer the core microbiome structure as comparable to the PE dataset. Conspicuously, the forward (R1) datasets inferred a higher number of taxa as compared to PE datasets for most of the amplicon regions, except V3V4. Thus, analyses of SE sequencing data, especially R1 reads, in environmental microbiome studies could ameliorate the problems arising on sample size of the study due to low quality reverse reads in the dataset. However, care must be taken while interpreting the microbiome structure as few taxa observed in the PE datasets were absent in the SE datasets. In conclusion, this study demonstrates the availability of choices in analyzing the amplicon data without having the need to remove samples with low quality reverse reads.

Tumor-Infiltrating B- and T-Cell Repertoire in Pancreatic Cancer Associated With Host and Tumor Features

Background

Infiltrating B and T cells have been observed in several tumor tissues, including pancreatic ductal adenocarcinoma (PDAC). The majority known PDAC risk factors point to a chronic inflammatory process leading to different forms of immunological infiltration. Understanding pancreatic tumor infiltration may lead to improved knowledge of this devastating disease.

Methods

We extracted the immunoglobulins (IGs) and T cell receptors (TCRs) from RNA-sequencing of 144 PDAC from TCGA and 180 pancreatic normal tissue from GTEx. We used Shannon entropy to find differences in IG/TCR diversity. We performed a clonotype analysis considering the IG clone definition (same V and J segments, same CDR3 length, and 90% nucleotide identity between CDR3s) to study differences among the tumor samples. Finally, we performed an association analysis to find host and tumor factors associated with the IG/TCR.

Results

PDAC presented a richer and more diverse IG and TCR infiltration than normal pancreatic tissue. A higher IG infiltration was present in heavy smokers and females and it was associated with better overall survival. In addition, specific IG clonotypes classified samples with better prognosis explaining 24% of the prognosis phenotypic variance. On the other hand, a larger TCR infiltration was present in patients with previous history of diabetes and was associated with lower nonantigen load.

Conclusions

Our findings support PDAC subtyping according to its immune repertoire landscape with a potential impact on the understanding of the inflammatory basis of PDAC risk factors as well as the design of treatment options and prognosis monitoring.

Learning sparse log-ratios for high-throughput sequencing data

MOTIVATION

The automatic discovery of sparse biomarkers that are associated with an outcome of interest is a central goal of bioinformatics. In the context of high-throughput sequencing (HTS) data, and compositional data (CoDa) more generally, an important class of biomarkers are the log-ratios between the input variables. However, identifying predictive log-ratio biomarkers from HTS data is a combinatorial optimization problem, which is computationally challenging. Existing methods are slow to run and scale poorly with the dimension of the input, which has limited their application to low- and moderate-dimensional metagenomic datasets.

RESULTS

Building on recent advances from the field of deep learning, we present CoDaCoRe, a novel learning algorithm that identifies sparse, interpretable and predictive log-ratio biomarkers. Our algorithm exploits a continuous relaxation to approximate the underlying combinatorial optimization problem. This relaxation can then be optimized efficiently using the modern ML toolbox, in particular, gradient descent. As a result, CoDaCoRe runs several orders of magnitude faster than competing methods, all while achieving state-of-the-art performance in terms of predictive accuracy and sparsity. We verify the outperformance of CoDaCoRe across a wide range of microbiome, metabolite and microRNA benchmark datasets, as well as a particularly high-dimensional dataset that is outright computationally intractable for existing sparse log-ratio selection methods.

AVAILABILITY AND IMPLEMENTATION

The CoDaCoRe package is available at https://github.com/egr95/R-codacore. Code and instructions for reproducing our results are available at https://github.com/cunningham-lab/codacore.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Opportunities and limits of combining microbiome and genome data for complex trait prediction

Background

Analysis and prediction of complex traits using microbiome data combined with host genomic information is a topic of utmost interest. However, numerous questions remain to be answered: how useful can the microbiome be for complex trait prediction? Are estimates of microbiability reliable? Can the underlying biological links between the host’s genome, microbiome, and phenome be recovered?

Methods

Here, we address these issues by (i) developing a novel simulation strategy that uses real microbiome and genotype data as inputs, and (ii) using variance-component approaches (Bayesian Reproducing Kernel Hilbert Space (RKHS) and Bayesian variable selection methods (Bayes C)) to quantify the proportion of phenotypic variance explained by the genome and the microbiome. The proposed simulation approach can mimic genetic links between the microbiome and genotype data by a permutation procedure that retains the distributional properties of the data.

Results

Using real genotype and rumen microbiota abundances from dairy cattle, simulation results suggest that microbiome data can significantly improve the accuracy of phenotype predictions, regardless of whether some microbiota abundances are under direct genetic control by the host or not. This improvement depends logically on the microbiome being stable over time. Overall, random-effects linear methods appear robust for variance components estimation, in spite of the typically highly leptokurtic distribution of microbiota abundances. The predictive performance of Bayes C was higher but more sensitive to the number of causative effects than RKHS. Accuracy with Bayes C depended, in part, on the number of microorganisms’ taxa that influence the phenotype.

Conclusions

While we conclude that, overall, genome-microbiome-links can be characterized using variance component estimates, we are less optimistic about the possibility of identifying the causative host genetic effects that affect microbiota abundances, which would require much larger sample sizes than are typically available for genome-microbiome-phenome studies. The R code to replicate the analyses is in https://github.com/miguelperezenciso/simubiome.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12711-021-00658-7.

How does the early life environment influence the oral microbiome and determine oral health outcomes in childhood?

The first 1000 days of life, from conception to 2 years, are a critical window for the influence of environmental exposures on the assembly of the oral microbiome, which is the precursor to dental caries (decay), one of the most prevalent microbially induced disorders worldwide. While it is known that the human microbiome is susceptible to environmental exposures, there is limited understanding of the impact of prenatal and early childhood exposures on the oral microbiome trajectory and oral health. A barrier has been the lack of technology to directly measure the foetal "exposome", which includes nutritional and toxic exposures crossing the placenta. Another barrier has been the lack of statistical methods to account for the high dimensional data generated by-omic assays. Through identifying which early life exposures influence the oral microbiome and modify oral health, these findings can be translated into interventions to reduce dental decay prevalence.

Unraveling the predictive role of temperature in the gut microbiota of the sea urchin Echinometra sp. EZ across spatial and temporal gradients

Shifts in microbial communities represent a rapid response mechanism for host organisms to respond to changes in environmental conditions. Therefore, they are likely to be important in assisting the acclimatization of hosts to seasonal temperature changes as well as to variation in temperatures across a species' range. The Persian/Arabian Gulf is the world's warmest sea, with large seasonal fluctuations in temperature (20℃ - 37℃) and is connected to the Gulf of Oman which experiences more typical oceanic conditions (<32℃ in the summer). This system is an informative model for understanding how symbiotic microbial assemblages respond to thermal variation across temporal and spatial scales. Here, we elucidate the role of temperature on the microbial gut community of the sea urchin Echinometra sp. EZ and identify microbial taxa that are tightly correlated with the thermal environment. We generated two independent datasets with a high degree of geographic and temporal resolution. The results show that microbial communities vary across thermally variable habitats, display temporal shifts that correlate with temperature, and can become more disperse as temperatures rise. The relative abundances of several ASVs significantly correlate with temperature in both independent datasets despite the >300 km distance between the furthest sites and the extreme seasonal variations. Notably, over 50% of the temperature predictive ASVs identified from the two datasets belonged to the family Vibrionaceae. Together, our results identify temperature as a robust predictor of community-level variation and highlight specific microbial taxa putatively involved in the response to thermal environment.

DisBalance: a platform to automatically build balance-based disease prediction models and discover microbial biomarkers from microbiome data

How best to utilize the microbial taxonomic abundances in regard to the prediction and explanation of human diseases remains appealing and challenging, and the relative nature of microbiome data necessitates a proper feature selection method to resolve the compositional problem. In this study, we developed an all-in-one platform to address a series of issues in microbiome-based human disease prediction and taxonomic biomarkers discovery. We prioritize the interpretation, runtime and classification accuracy of the distal discriminative balances analysis (DBA-distal) method in selecting a set of distal discriminative balances, and develop DisBalance, a comprehensive platform, to integrate and streamline the workflows of disease model building, disease risk prediction and disease-related biomarker discovery for microbiome-based binary classifications. DisBalance allows the de novo model-building and disease risk prediction in a very fast and convenient way. To facilitate the model-driven and knowledge-driven discoveries, DisBalance dedicates multiple strategies for the mining of microbial biomarkers. The independent validation of the models constructed by the DisBalance pipeline is performed on seven microbiome datasets from the original article of DBA-distal. The implementation of the DisBalance platform is demonstrated by a complete analysis of a shotgun metagenomic dataset of Ulcerative Colitis (UC). As a free and open-source, DisBlance can be accessed at http://lab.malab.cn/soft/DisBalance. The source code and demo data for Disbalance are available at https://github.com/yangfenglong/DisBalance.