ABSTRACT
Time series studies of microbiome interventions provide valuable data about microbial ecosystem structure. Unfortunately, existing models of microbial community dynamics have limited temporal memory and expressivity, relying on Markov or linearity assumptions. To address this, we introduce a new class of models based on transfer functions. These models learn impulse responses, capturing the potentially delayed effects of environmental changes on the microbial community. This allows us to simulate trajectories under hypothetical interventions and select significantly perturbed taxa with False Discovery Rate guarantees. Through simulations, we show that our approach effectively reduces forecasting errors compared to strong baselines and accurately pinpoints taxa of interest. Our case studies highlight the interpretability of the resulting differential response trajectories. An R package, mbtransfer, and notebooks to replicate the simulation and case studies are provided.
Subject(s)
Computational Biology , Microbiota , Microbiota/physiology , Humans , Computational Biology/methods , Computer Simulation , Models, Biological , Software , Markov ChainsABSTRACT
Diverse and non-Lactobacillus-dominated vaginal microbial communities are associated with adverse health outcomes such as preterm birth and the acquisition of sexually transmitted infections. Despite the importance of recognizing and understanding the key risk-associated features of these communities, their heterogeneous structure and properties remain ill-defined. Clustering approaches are commonly used to characterize vaginal communities, but they lack sensitivity and robustness in resolving substructures and revealing transitions between potential sub-communities. Here, we address this need with an approach based on mixed membership topic models. Using longitudinal data from cohorts of pregnant and non-pregnant study participants, we show that topic models more accurately describe sample composition, longitudinal changes, and better predict the loss of Lactobacillus dominance. We identify several non-Lactobacillus-dominated sub-communities common to both cohorts and independent of reproductive status. In non-pregnant individuals, we find that the menstrual cycle modulates transitions between and within sub-communities, as well as the concentrations of half of the cytokines and 18% of metabolites. Overall, our analyses based on mixed membership models reveal substructures of vaginal ecosystems which may have important clinical and biological associations.
Subject(s)
Microbiota , Premature Birth , Pregnancy , Female , Infant, Newborn , Humans , Vagina , Lactobacillus/metabolism , Menstrual Cycle , RNA, Ribosomal, 16SSubject(s)
Computational Biology , Single-Cell Analysis , Humans , Proteomics , Software , TranscriptomeABSTRACT
High throughput sequencing (HTS)-based technology enables identifying and quantifying non-culturable microbial organisms in all environments. Microbial sequences have enhanced our understanding of the human microbiome, the soil and plant environment, and the marine environment. All molecular microbial data pose statistical challenges due to contamination sequences from reagents, batch effects, unequal sampling, and undetected taxa. Technical biases and heteroscedasticity have the strongest effects, but different strains across subjects and environments also make direct differential abundance testing unwieldy. We provide an introduction to a few statistical tools that can overcome some of these difficulties and demonstrate those tools on an example. We show how standard statistical methods, such as simple hierarchical mixture and topic models, can facilitate inferences on latent microbial communities. We also review some nonparametric Bayesian approaches that combine visualization and uncertainty quantification. The intersection of molecular microbial biology and statistics is an exciting new venue. Finally, we list some of the important open problems that would benefit from more careful statistical method development.
ABSTRACT
Our work focuses on the stability, resilience, and response to perturbation of the bacterial communities in the human gut. Informative flash flood-like disturbances that eliminate most gastrointestinal biomass can be induced using a clinically-relevant iso-osmotic agent. We designed and executed such a disturbance in human volunteers using a dense longitudinal sampling scheme extending before and after induced diarrhea. This experiment has enabled a careful multidomain analysis of a controlled perturbation of the human gut microbiota with a new level of resolution. These new longitudinal multidomain data were analyzed using recently developed statistical methods that demonstrate improvements over current practices. By imposing sparsity constraints we have enhanced the interpretability of the analyses and by employing a new adaptive generalized principal components analysis, incorporated modulated phylogenetic information and enhanced interpretation through scoring of the portions of the tree most influenced by the perturbation. Our analyses leverage the taxa-sample duality in the data to show how the gut microbiota recovers following this perturbation. Through a holistic approach that integrates phylogenetic, metagenomic and abundance information, we elucidate patterns of taxonomic and functional change that characterize the community recovery process across individuals. We provide complete code and illustrations of new sparse statistical methods for high-dimensional, longitudinal multidomain data that provide greater interpretability than existing methods.
Subject(s)
Gastrointestinal Microbiome/genetics , Metagenome/genetics , Metagenomics/methods , Models, Biological , Adult , DNA, Bacterial/analysis , DNA, Bacterial/genetics , Diarrhea , Female , Humans , Longitudinal Studies , Male , Middle Aged , Principal Component Analysis , RNA, Ribosomal, 16S/genetics , Young AdultABSTRACT
Preterm birth (PTB) is the leading cause of neonatal morbidity and mortality. Previous studies have suggested that the maternal vaginal microbiota contributes to the pathophysiology of PTB, but conflicting results in recent years have raised doubts. We conducted a study of PTB compared with term birth in two cohorts of pregnant women: one predominantly Caucasian (n = 39) at low risk for PTB, the second predominantly African American and at high-risk (n = 96). We profiled the taxonomic composition of 2,179 vaginal swabs collected prospectively and weekly during gestation using 16S rRNA gene sequencing. Previously proposed associations between PTB and lower Lactobacillus and higher Gardnerella abundances replicated in the low-risk cohort, but not in the high-risk cohort. High-resolution bioinformatics enabled taxonomic assignment to the species and subspecies levels, revealing that Lactobacillus crispatus was associated with low risk of PTB in both cohorts, while Lactobacillus iners was not, and that a subspecies clade of Gardnerella vaginalis explained the genus association with PTB. Patterns of cooccurrence between L. crispatus and Gardnerella were highly exclusive, while Gardnerella and L. iners often coexisted at high frequencies. We argue that the vaginal microbiota is better represented by the quantitative frequencies of these key taxa than by classifying communities into five community state types. Our findings extend and corroborate the association between the vaginal microbiota and PTB, demonstrate the benefits of high-resolution statistical bioinformatics in clinical microbiome studies, and suggest that previous conflicting results may reflect the different risk profile of women of black race.
Subject(s)
Premature Birth/microbiology , Vagina/microbiology , Adult , Black or African American , Case-Control Studies , Cohort Studies , DNA Replication , Female , Gardnerella vaginalis/classification , Humans , Lactobacillus/classification , Microbiota/genetics , Microbiota/immunology , Pregnancy , Premature Birth/etiology , RNA, Ribosomal, 16S/genetics , United States/epidemiology , White PeopleABSTRACT
OBJECTIVE: In a previous study, we demonstrated the therapeutic efficacy of a subcutaneous injection of GK1 peptide in a melanoma mouse model, effectively increasing the mean survival time by 42.58%, delaying tumor growth, and increasing intratumoral necrosis compared with the control. As a first approach to investigate the anti-melanoma effect of GK1, this study was carried out to determine the hematological effects along with both serum and lung cytokine profiles in a melanoma lung metastatic model. MATERIALS AND METHODS: Thirteen C57BL6 female mice were transfected in the lateral tail vein with 2×10(5) B16-F0 melanoma cells. After 7 days, mice were separated in two different groups and treatments were initiated (day 0): The GK1-treated group (seven mice) were injected every 5 days intravenously with GK1 (10 µg) in the lateral tail vein, and the control group (six mice) were injected every 5 days with intravenous saline solution. Blood samples were collected every 5 days from day 0; tumor samples were obtained for cytokine measurements on the day of sacrifice. RESULTS: In the peripheral blood, mice treated with GK1 presented a statistically significant decrease in IFN-γ (p<0.05), and lymphocytes tended to be lower compared with the control mice (p=0.06). Lung metastatic analysis demonstrated a significant increase in IFN-γ and IL-12p70 (p<0.05); a significant decrease in IL-17, IL-4, IL-22, IL-23, and IL-12p40 (p<0.05); and a marginal decrease in IL-1ß (p=0.07) compared with the control. DISCUSSION: Our results suggest that an intratumoral increase of cytokines with antitumor activity along with an intratumoral decrease of cytokines with protumor activity could explain, in part, the anti-melanoma effects of GK1 in a lung metastatic melanoma mouse model. Further studies must be performed to elucidate the precise mechanisms of action for GK1 peptide against melanoma, and their eventual application in humans.