SpotSweeper: spatially-aware quality control for spatial transcriptomics.

Quality control (QC) is a crucial step to ensure the reliability and accuracy of the data obtained from RNA sequencing experiments, including spatially-resolved transcriptomics (SRT). Existing QC approaches for SRT that have been adopted from single-nucleus RNA sequencing (snRNA-seq) methods are confounded by spatial biology and are inappropriate for SRT data. In addition, no methods currently exist for identifying histological tissue artifacts unique to SRT. Here, we introduce SpotSweeper, spatially-aware QC methods for identifying local outliers and regional artifacts in SRT. SpotSweeper evaluates the quality of individual spots relative to their local neighborhood, thus minimizing bias due to biological heterogeneity, and uses multiscale methods to detect regional artifacts. Using SpotSweeper on publicly available data, we identified a consistent set of Visium barcodes/spots as systematically low quality and demonstrate that SpotSweeper accurately identifies two distinct types of regional artifacts, resulting in improved downstream clustering and marker gene detection for spatial domains.

A data-driven single-cell and spatial transcriptomic map of the human prefrontal cortex.

The molecular organization of the human neocortex historically has been studied in the context of its histological layers. However, emerging spatial transcriptomic technologies have enabled unbiased identification of transcriptionally defined spatial domains that move beyond classic cytoarchitecture. We used the Visium spatial gene expression platform to generate a data-driven molecular neuroanatomical atlas across the anterior-posterior axis of the human dorsolateral prefrontal cortex. Integration with paired single-nucleus RNA-sequencing data revealed distinct cell type compositions and cell-cell interactions across spatial domains. Using PsychENCODE and publicly available data, we mapped the enrichment of cell types and genes associated with neuropsychiatric disorders to discrete spatial domains.

Neighborhood Disadvantage and Risk of Heart Failure: The Reasons for Geographic and Racial Differences in Stroke (REGARDS) Study.

BACKGROUND: Heart failure (HF) affects >6 million US adults, with recent increases in HF hospitalizations. We aimed to investigate the association between neighborhood disadvantage and incident HF events and potential differences by diabetes status. METHODS: We included 23 645 participants from the REGARDS study (Reasons for Geographic and Racial Differences in Stroke), a prospective cohort of Black and White adults aged ≥45 years living in the continental United States (baseline 2005-2007). Neighborhood disadvantage was assessed using a Z score of 6 census tract variables (2000 US Census) and categorized as quartiles. Incident HF hospitalizations or HF-related deaths through 2017 were adjudicated. Multivariable-adjusted Cox regression was used to examine the association between neighborhood disadvantage and incident HF. Heterogeneity by diabetes was assessed using an interaction term. RESULTS: The mean age was 64.4 years, 39.5% were Black adults, 54.9% females, and 18.8% had diabetes. During a median follow-up of 10.7 years, there were 1125 incident HF events with an incidence rate of 3.3 (quartile 1), 4.7 (quartile 2), 5.2 (quartile 3), and 6.0 (quartile 4) per 1000 person-years. Compared to adults living in the most advantaged neighborhoods (quartile 1), those living in neighborhoods in quartiles 2, 3, and 4 (most disadvantaged) had 1.30 (95% CI, 1.06-1.60), 1.36 (95% CI, 1.11-1.66), and 1.45 (95% CI, 1.18-1.79) times greater hazard of incident HF even after accounting for known confounders. This association did not significantly differ by diabetes status (interaction P=0.59). For adults with diabetes, the adjusted incident HF hazards comparing those in quartile 4 versus quartile 1 was 1.34 (95% CI, 0.92-1.96), and it was 1.50 (95% CI, 1.16-1.94) for adults without diabetes. CONCLUSIONS: In this large contemporaneous prospective cohort, neighborhood disadvantage was associated with an increased risk of incident HF events. This increase in HF risk did not differ by diabetes status. Addressing social, economic, and structural factors at the neighborhood level may impact HF prevention.

escheR: unified multi-dimensional visualizations with Gestalt principles.

Summary: The creation of effective visualizations is a fundamental component of data analysis. In biomedical research, new challenges are emerging to visualize multi-dimensional data in a 2D space, but current data visualization tools have limited capabilities. To address this problem, we leverage Gestalt principles to improve the design and interpretability of multi-dimensional data in 2D data visualizations, layering aesthetics to display multiple variables. The proposed visualization can be applied to spatially-resolved transcriptomics data, but also broadly to data visualized in 2D space, such as embedding visualizations. We provide an open source R package escheR, which is built off of the state-of-the-art ggplot2 visualization framework and can be seamlessly integrated into genomics toolboxes and workflows. Availability and implementation: The open source R package escheR is freely available on Bioconductor (https://bioconductor.org/packages/escheR).

Gut microbiota impacts bone via Bacteroides vulgatus-valeric acid-related pathways.

Although the gut microbiota has been reported to influence osteoporosis risk, the individual species involved, and underlying mechanisms, remain largely unknown. We performed integrative analyses in a Chinese cohort of peri-/post-menopausal women with metagenomics/targeted metabolomics/whole-genome sequencing to identify novel microbiome-related biomarkers for bone health. Bacteroides vulgatus was found to be negatively associated with bone mineral density (BMD), which was validated in US white people. Serum valeric acid (VA), a microbiota derived metabolite, was positively associated with BMD and causally downregulated by B. vulgatus. Ovariectomized mice fed B. vulgatus demonstrated increased bone resorption and poorer bone micro-structure, while those fed VA demonstrated reduced bone resorption and better bone micro-structure. VA suppressed RELA protein production (pro-inflammatory), and enhanced IL10 mRNA expression (anti-inflammatory), leading to suppressed maturation of osteoclast-like cells and enhanced maturation of osteoblasts in vitro. The findings suggest that B. vulgatus and VA may represent promising targets for osteoporosis prevention/treatment.

escheR: Unified multi-dimensional visualizations with Gestalt principles.

The creation of effective visualizations is a fundamental component of data analysis. In biomedical research, new challenges are emerging to visualize multi-dimensional data in a 2D space, but current data visualization tools have limited capabilities. To address this problem, we leverage Gestalt principles to improve the design and interpretability of multi-dimensional data in 2D data visualizations, layering aesthetics to display multiple variables. The proposed visualization can be applied to spatially-resolved transcriptomics data, but also broadly to data visualized in 2D space, such as embedding visualizations. We provide an open source R package escheR, which is built off of the state-of-the-art ggplot2 visualization framework and can be seamlessly integrated into genomics toolboxes and workflows.

Integrated single cell and unsupervised spatial transcriptomic analysis defines molecular anatomy of the human dorsolateral prefrontal cortex.

Generation of a molecular neuroanatomical map of the human prefrontal cortex reveals novel spatial domains and cell-cell interactions relevant for psychiatric disease. The molecular organization of the human neocortex has been historically studied in the context of its histological layers. However, emerging spatial transcriptomic technologies have enabled unbiased identification of transcriptionally-defined spatial domains that move beyond classic cytoarchitecture. Here we used the Visium spatial gene expression platform to generate a data-driven molecular neuroanatomical atlas across the anterior-posterior axis of the human dorsolateral prefrontal cortex (DLPFC). Integration with paired single nucleus RNA-sequencing data revealed distinct cell type compositions and cell-cell interactions across spatial domains. Using PsychENCODE and publicly available data, we map the enrichment of cell types and genes associated with neuropsychiatric disorders to discrete spatial domains. Finally, we provide resources for the scientific community to explore these integrated spatial and single cell datasets at research.libd.org/spatialDLPFC/.

D-dimer and the risk of hypertension: The REasons for Geographic And Racial Differences in Stroke Cohort Study.

Background: Reasons for increased risk of hypertension in Black compared with White people are only partly understood. D-dimer, a thrombo-inflammatory marker higher in Black individuals, is also higher in people with hypertension. However, the impact of D-dimer on racial disparities in risk of incident hypertension has not been studied. Objectives: To assess whether D-dimer is associated with the risk of incident hypertension, whether the association between D-dimer and the risk of incident hypertension differs by race, and whether the biology reflected by D-dimer explains racial disparities in the risk of incident hypertension. Methods: This study included 1867 participants in the REasons for Geographic And Racial Differences in Stroke cohort study without baseline hypertension and with a second visit 9.4 years after baseline. Risk ratios of incident hypertension by baseline D-dimer level were estimated, a D-dimer-by-race interaction was tested, and the mediating effect of D-dimer (which represents underlying biological processes) on the association of race and hypertension risk was assessed. Results: The risk of incident hypertension was 47% higher in persons in the top quartile than in those in the bottom quartile of D-dimer (risk ratio [RR]: 1.47; 95% CI: 1.23-1.76). The association was partly attenuated after adjusting for sociodemographic and adiposity-related risk factors (RR: 1.22; 95% CI: 1.02-1.47). The association of D-dimer and hypertension did not differ by race, and D-dimer did not attenuate the racial difference in the risk of incident hypertension. Conclusion: D-dimer concentration reflects pathophysiology related to the development of hypertension. Specific mechanisms require further study and may involve adiposity.

Plasma Metabolites Link Dietary Patterns to Stroke Risk.

OBJECTIVE: While dietary intake is linked to stroke risk, surrogate markers that could inform personalized dietary interventions are lacking. We identified metabolites associated with diet patterns and incident stroke in a nested cohort from the REasons for Geographic and Racial Differences in Stroke (REGARDS) study. METHODS: Levels of 162 metabolites were measured in baseline plasma from stroke cases (n = 1,198) and random controls (n = 904). We examined associations between metabolites and a plant-based diet pattern previously linked to reduced stroke risk in REGARDS. Secondary analyses included 3 additional stroke-associated diet patterns: a Mediterranean, Dietary Approaches to Stop Hypertension (DASH), and Southern diet. Metabolites were tested using Cox proportional hazards models with incident stroke as the outcome. Replication was performed in the Jackson Heart Study (JHS). Inverse odds ratio-weighted mediation was used to determine whether metabolites mediated the association between a plant-based diet and stroke risk. RESULTS: Metabolites associated with a plant-based diet included the gut metabolite indole-3-propionic acid (ß = 0.23, 95% confidence interval [CI] [0.14, 0.33], p = 1.14 × 10-6 ), guanosine (ß = -0.13, 95% CI [-0.19, -0.07], p = 6.48 × 10-5 ), gluconic acid (ß = -0.11, 95% CI [-0.18, -0.04], p = 2.06 × 10-3 ), and C7 carnitine (ß = -0.16, 95% CI [-0.24, -0.09], p = 4.14 × 10-5 ). All of these metabolites were associated with both additional diet patterns and altered stroke risk. Mediation analyses identified guanosine (32.6% mediation, p = 1.51 × 10-3 ), gluconic acid (35.7%, p = 2.28 × 10-3 ), and C7 carnitine (26.2%, p = 1.88 × 10-2 ) as mediators linking a plant-based diet to reduced stroke risk. INTERPRETATION: A subset of diet-related metabolites are associated with risk of stroke. These metabolites could serve as surrogate markers that inform dietary interventions. ANN NEUROL 2023;93:500-510.

Spike-and-slab least absolute shrinkage and selection operator generalized additive models and scalable algorithms for high-dimensional data analysis.

There are proposals that extend the classical generalized additive models (GAMs) to accommodate high-dimensional data ( p ≫ n $$ p\gg n $$ ) using group sparse regularization. However, the sparse regularization may induce excess shrinkage when estimating smooth functions, damaging predictive performance. Moreover, most of these GAMs consider an "all-in-all-out" approach for functional selection, rendering them difficult to answer if nonlinear effects are necessary. While some Bayesian models can address these shortcomings, using Markov chain Monte Carlo algorithms for model fitting creates a new challenge, scalability. Hence, we propose Bayesian hierarchical generalized additive models as a solution: we consider the smoothing penalty for proper shrinkage of curve interpolation via reparameterization. A novel two-part spike-and-slab LASSO prior for smooth functions is developed to address the sparsity of signals while providing extra flexibility to select the linear or nonlinear components of smooth functions. A scalable and deterministic algorithm, EM-Coordinate Descent, is implemented in an open-source R package BHAM. Simulation studies and metabolomics data analyses demonstrate improved predictive and computational performance against state-of-the-art models. Functional selection performance suggests trade-offs exist regarding the effect hierarchy assumption.

C-reactive Protein and Racial Differences in Type 2 Diabetes Incidence: The REGARDS Study.

CONTEXT: Black adults experience more type 2 diabetes mellitus and higher inflammatory markers, including C-reactive protein (CRP), than White adults. Inflammatory markers are associated with risk of incident diabetes but the impact of inflammation on racial differences in incident diabetes is unknown. OBJECTIVE: We assessed whether CRP mediated the Black-White incident diabetes disparity. METHODS: The REasons for Geographic And Racial Differences in Stroke (REGARDS) study enrolled 30 239 US Black and White adults aged ≥45 years in 2003-2007 with a second visit approximately 10 years later. Among participants without baseline diabetes, adjusted sex- and race-stratified risk ratios for incident diabetes at the second visit by CRP level were calculated using modified Poisson regression. Inverse odds weighting estimated the percent mediation of the racial disparity by CRP. RESULTS: Of 11 073 participants without baseline diabetes (33% Black, 67% White), 1389 (12.5%) developed diabetes. Black participants had higher CRP at baseline and greater incident diabetes than White participants. Relative to CRP < 3 mg/L, CRP ≥ 3 mg/L was associated with greater risk of diabetes in all race-sex strata. Black participants had higher risk of diabetes at CRP < 3 mg/L, but not at CRP ≥ 3 mg/L. In women, CRP mediated 10.0% of the racial difference in incident diabetes. This mediation was not seen in men. CONCLUSION: Higher CRP is a risk factor for incident diabetes, but the excess burden of diabetes in Black adults was only seen in those with lower CRP, suggesting that inflammation is unlikely to be the main driver of this racial disparity.

Biomarkers as MEDiators of racial disparities in risk factors (BioMedioR): Rationale, study design, and statistical considerations.

PURPOSE: Relative to White adults, Black adults have a substantially higher prevalence of hypertension and diabetes, both key risk factors for stroke, cardiovascular disease, cognitive impairment, and dementia. Blood biomarkers have shown promise in identifying contributors to racial disparities in many chronic diseases. METHODS: We outline the study design and related statistical considerations for a nested cohort study, the Biomarker Mediators of Racial Disparities in Risk Factors (BioMedioR) study, within the 30,239-person biracial REasons for Geographic And Racial Differences in Stroke (REGARDS) study (2003-present). Selected biomarkers will be assessed for contributions to racial disparities in risk factor development over median 9.4 years of follow-up, with initial focus on hypertension, and diabetes. Here we outline study design decisions and statistical considerations for the sampling of 4,400 BioMedioR participants. RESULTS: The population for biomarker assessment was selected using a random sample study design balanced across race and sex to provide the optimal opportunity to describe association of biomarkers with the development of hypertension and diabetes. Descriptive characteristics of the BioMedioR sample and analytic plans are provided for this nested cohort study. CONCLUSIONS: This nested biomarker study will examine pathways with the target to help explain racial differences in hypertension and diabetes incidence.

Does the Association Between Hemoglobin A1c and Risk of Cardiovascular Events Vary by Residential Segregation? The REasons for Geographic And Racial Differences in Stroke (REGARDS) Study.

OBJECTIVE: To examine if the association between higher A1C and risk of cardiovascular disease (CVD) among adults with and without diabetes is modified by racial residential segregation. RESEARCH DESIGN AND METHODS: The study used a case-cohort design, which included a random sample of 2,136 participants at baseline and 1,248 participants with incident CVD (i.e., stroke, coronary heart disease [CHD], and fatal CHD during 7-year follow-up) selected from 30,239 REasons for Geographic And Racial Differences in Stroke (REGARDS) study participants originally assessed between 2003 and 2007. The relationship of A1C with incident CVD, stratified by baseline diabetes status, was assessed using Cox proportional hazards models adjusting for demographics, CVD risk factors, and socioeconomic status. Effect modification by census tract-level residential segregation indices (dissimilarity, interaction, and isolation) was assessed using interaction terms. RESULTS: The mean age of participants in the random sample was 64.2 years, with 44% African American, 59% female, and 19% with diabetes. In multivariable models, A1C was not associated with CVD risk among those without diabetes (hazard ratio [HR] per 1% [11 mmol/mol] increase, 0.94 [95% CI 0.76-1.16]). However, A1C was associated with an increased risk of CVD (HR per 1% increase, 1.23 [95% CI 1.08-1.40]) among those with diabetes. This A1C-CVD association was modified by the dissimilarity (P < 0.001) and interaction (P = 0.001) indices. The risk of CVD was increased at A1C levels between 7 and 9% (53-75 mmol/mol) for those in areas with higher residential segregation (i.e., lower interaction index). In race-stratified analyses, there was a more pronounced modifying effect of residential segregation among African American participants with diabetes. CONCLUSIONS: Higher A1C was associated with increased CVD risk among individuals with diabetes, and this relationship was more pronounced at higher levels of residential segregation among African American adults. Additional research on how structural determinants like segregation may modify health effects is needed.

C-Reactive Protein and Incident Hypertension in Black and White Americans in the REasons for Geographic And Racial Differences in Stroke (REGARDS) Cohort Study.

BACKGROUND: More inflammation is associated with greater risk incident hypertension, and Black United States (US) adults have excess burden of hypertension. We investigated whether increased inflammation as quantified by higher C-reactive protein (CRP) explains the excess incidence in hypertension experienced by Black US adults. METHODS: We included 6,548 Black and White REasons for Geographic and Racial Differences in Stroke (REGARDS) participants without hypertension at baseline (2003-2007) who attended a second visit (2013-2016). Sex-stratified risk ratios (RRs) for incident hypertension at the second exam in Black compared to White individuals were estimated using Poisson regression adjusted for groups of factors known to partially explain the Black-White differences in incident hypertension. We calculated the percent mediation by CRP of the racial difference in hypertension. RESULTS: Baseline CRP was higher in Black participants. The Black-White RR for incident hypertension in the minimally adjusted model was 1.33 (95% confidence interval 1.22, 1.44) for males and 1.15 (1.04, 1.27) for females. CRP mediated 6.6% (95% confidence interval 2.7, 11.3%) of this association in females and 19.7% (9.8, 33.2%) in males. In females, CRP no longer mediated the Black-White RR in a model including waist circumference and body mass index, while in males the Black-White difference was fully attenuated in models including income, education and dietary patterns. CONCLUSIONS: Elevated CRP attenuated a portion of the unadjusted excess risk of hypertension in Black adults, but this excess risk was attenuated when controlling for measures of obesity in females and diet and socioeconomic factors in males. Inflammation related to these risk factors might explain part of the Black-White disparity in hypertension.

Zero-Inflated gaussian mixed models for analyzing longitudinal microbiome data.

MOTIVATION: The human microbiome is variable and dynamic in nature. Longitudinal studies could explain the mechanisms in maintaining the microbiome in health or causing dysbiosis in disease. However, it remains challenging to properly analyze the longitudinal microbiome data from either 16S rRNA or metagenome shotgun sequencing studies, output as proportions or counts. Most microbiome data are sparse, requiring statistical models to handle zero-inflation. Moreover, longitudinal design induces correlation among the samples and thus further complicates the analysis and interpretation of the microbiome data. RESULTS: In this article, we propose zero-inflated Gaussian mixed models (ZIGMMs) to analyze longitudinal microbiome data. ZIGMMs is a robust and flexible method which can be applicable for longitudinal microbiome proportion data or count data generated with either 16S rRNA or shotgun sequencing technologies. It can include various types of fixed effects and random effects and account for various within-subject correlation structures, and can effectively handle zero-inflation. We developed an efficient Expectation-Maximization (EM) algorithm to fit the ZIGMMs by taking advantage of the standard procedure for fitting linear mixed models. We demonstrate the computational efficiency of our EM algorithm by comparing with two other zero-inflated methods. We show that ZIGMMs outperform the previously used linear mixed models (LMMs), negative binomial mixed models (NBMMs) and zero-inflated Beta regression mixed model (ZIBR) in detecting associated effects in longitudinal microbiome data through extensive simulations. We also apply our method to two public longitudinal microbiome datasets and compare with LMMs and NBMMs in detecting dynamic effects of associated taxa.

Bayesian hierarchical negative binomial models for multivariable analyses with applications to human microbiome count data.

The analyses of large volumes of metagenomic data extracted from aggregate populations of microscopic organisms residing on and in the human body are advancing contemporary understandings of the integrated participation of microbes in human health and disease. Next generation sequencing technology facilitates said analyses in terms of diversity, community composition, and differential abundance by filtering and binning microbial 16S rRNA genes extracted from human tissues into operational taxonomic units. However, current statistical tools restrict study designs to investigations of limited numbers of host characteristics mediated by limited numbers of samples potentially yielding a loss of relevant information. This paper presents a Bayesian hierarchical negative binomial model as an efficient technique capable of compensating for multivariable sets including tens or hundreds of host characteristics as covariates further expanding analyses of human microbiome count data. Simulation studies reveal that the Bayesian hierarchical negative binomial model provides a desirable strategy by often outperforming three competing negative binomial model in terms of type I error while simultaneously maintaining consistent power. An application of the Bayesian hierarchical negative binomial model using subsets of the open data published by the American Gut Project demonstrates an ability to identify operational taxonomic units significantly differentiable among persons diagnosed by a medical professional with either inflammatory bowel disease or irritable bowel syndrome that are consistent with contemporary gastrointestinal literature.

WNT pathway signaling is associated with microvascular injury and predicts kidney transplant failure.

Microvascular injury is associated with accelerated kidney transplant dysfunction and allograft failure. Molecular pathology can identify new mechanisms of microvascular injury while improving on the diagnostic and prognostic capabilities of traditional histology. We conducted a case-control study of archived kidney biopsy specimens stored up to 10 years with microvascular injury (n = 50) compared with biopsy specimens without histologic injury (n = 45) from patients of similar age, race, and sex. We measured WNT gene expression with a multiplex quantification platform by using digital barcoding, given the importance of WNT reactivation to the response to wounding in the kidney microvasculature and other compartments. Of 210 genes from a commercial WNT panel, 71 were associated with microvascular injury and 79 were associated with allograft failure, with considerable overlap of genes between each set. Molecular pathology identified 46 biopsy specimens with molecular evidence of microvascular injury; 18 (39%) were either C4d negative, donor-specific antibody negative, or had no microvascular injury by histology. The majority of cases with molecular evidence of microvascular injury had poor long-term outcomes. We identified novel WNT pathway genes associated with microvascular injury and allograft failure in residual clinical biopsy specimens obtained up to 10 years earlier. Further mechanistic studies may identify the WNT pathway as a new diagnostic and therapeutic target.

Gsslasso Cox: a Bayesian hierarchical model for predicting survival and detecting associated genes by incorporating pathway information.

BACKGROUND: Group structures among genes encoded in functional relationships or biological pathways are valuable and unique features in large-scale molecular data for survival analysis. However, most of previous approaches for molecular data analysis ignore such group structures. It is desirable to develop powerful analytic methods for incorporating valuable pathway information for predicting disease survival outcomes and detecting associated genes. RESULTS: We here propose a Bayesian hierarchical Cox survival model, called the group spike-and-slab lasso Cox (gsslasso Cox), for predicting disease survival outcomes and detecting associated genes by incorporating group structures of biological pathways. Our hierarchical model employs a novel prior on the coefficients of genes, i.e., the group spike-and-slab double-exponential distribution, to integrate group structures and to adaptively shrink the effects of genes. We have developed a fast and stable deterministic algorithm to fit the proposed models. We performed extensive simulation studies to assess the model fitting properties and the prognostic performance of the proposed method, and also applied our method to analyze three cancer data sets. CONCLUSIONS: Both the theoretical and empirical studies show that the proposed method can induce weaker shrinkage on predictors in an active pathway, thereby incorporating the biological similarity of genes within a same pathway into the hierarchical modeling. Compared with several existing methods, the proposed method can more accurately estimate gene effects and can better predict survival outcomes. For the three cancer data sets, the results show that the proposed method generates more powerful models for survival prediction and detecting associated genes. The method has been implemented in a freely available R package BhGLM at https://github.com/nyiuab/BhGLM .

BhGLM: Bayesian hierarchical GLMs and survival models, with applications to genomics and epidemiology.

SUMMARY: BhGLM is a freely available R package that implements Bayesian hierarchical modeling for high-dimensional clinical and genomic data. It consists of functions for setting up various Bayesian hierarchical models, including generalized linear models (GLMs) and Cox survival models, with four types of prior distributions for coefficients, i.e. double-exponential, Student-t, mixture double-exponential and mixture Student-t. These functions adapt fast and stable algorithms to estimate parameters. BhGLM also provides functions for summarizing results numerically and graphically and for evaluating predictive values. The package is particularly useful for analyzing large-scale molecular data, i.e. detecting disease-associated variables and predicting disease outcomes. We here describe the models, algorithms and associated features implemented in BhGLM. AVAILABILITY AND IMPLEMENTATION: The package is freely available from the public GitHub repository, https://github.com/nyiuab/BhGLM.

Negative Binomial Mixed Models for Analyzing Longitudinal Microbiome Data.

The metagenomics sequencing data provide valuable resources for investigating the associations between the microbiome and host environmental/clinical factors and the dynamic changes of microbial abundance over time. The distinct properties of microbiome measurements include varied total sequence reads across samples, over-dispersion and zero-inflation. Additionally, microbiome studies usually collect samples longitudinally, which introduces time-dependent and correlation structures among the samples and thus further complicates the analysis and interpretation of microbiome count data. In this article, we propose negative binomial mixed models (NBMMs) for longitudinal microbiome studies. The proposed NBMMs can efficiently handle over-dispersion and varying total reads, and can account for the dynamic trend and correlation among longitudinal samples. We develop an efficient and stable algorithm to fit the NBMMs. We evaluate and demonstrate the NBMMs method via extensive simulation studies and application to a longitudinal microbiome data. The results show that the proposed method has desirable properties and outperform the previously used methods in terms of flexible framework for modeling correlation structures and detecting dynamic effects. We have developed an R package NBZIMM to implement the proposed method, which is freely available from the public GitHub repository http://github.com//nyiuab//NBZIMM and provides a useful tool for analyzing longitudinal microbiome data.