Multigroup analysis of compositions of microbiomes with covariate adjustments and repeated measures.

Microbiome differential abundance analysis methods for two groups are well-established in the literature. However, many microbiome studies involve more than two groups, sometimes even ordered groups such as stages of a disease, and require different types of comparison. Standard pairwise comparisons are inefficient in terms of power and false discovery rates. In this Article, we propose a general framework, ANCOM-BC2, for performing a wide range of multigroup analyses with covariate adjustments and repeated measures. We illustrate our methodology through two real datasets. The first example explores the effects of aridity on the soil microbiome, and the second example investigates the effects of surgical interventions on the microbiome of patients with inflammatory bowel disease.

Association between lipid-A-producing oral bacteria of different potency and fractional exhaled nitric oxide in a Norwegian population-based adult cohort.

BACKGROUND: Lipid A is the primary immunostimulatory part of the lipopolysaccharide (LPS) molecule. The inflammatory response of LPS varies and depends upon the number of acyl chains and phosphate groups in lipid A which is specific for a bacterial species or strain. Traditional LPS quantification assays cannot distinguish between the acylation degree of lipid A molecules, and therefore little is known about how bacteria with different inflammation-inducing potencies affect fractional exhaled nitric oxide (FeNO). We aimed to explore the association between pro-inflammatory hexa- and less inflammatory penta-acylated LPS-producing oral bacteria and FeNO as a marker of airway inflammation. METHODS: We used data from a population-based adult cohort from Norway (n = 477), a study center of the RHINESSA multi-center generation study. We applied statistical methods on the bacterial community- (prediction with MiRKAT) and genus-level (differential abundance analysis with ANCOM-BC) to investigate the association between the oral microbiota composition and FeNO. RESULTS: We found the overall composition to be significantly associated with increasing FeNO levels independent of covariate adjustment, and abundances of 27 bacterial genera to differ in individuals with high FeNO vs. low FeNO levels. Hexa- and penta-acylated LPS producers made up 2.4% and 40.8% of the oral bacterial genera, respectively. The Bray-Curtis dissimilarity within hexa- and penta-acylated LPS-producing oral bacteria was associated with increasing FeNO levels independent of covariate adjustment. A few single penta-acylated LPS producers were more abundant in individuals with low FeNO vs. high FeNO, while hexa-acylated LPS producers were found not to be enriched. CONCLUSIONS: In a population-based adult cohort, FeNO was observed to be associated with the overall oral bacterial community composition. The effect of hexa- and penta-acylated LPS-producing oral bacteria was overall significant when focusing on Bray-Curtis dissimilarity within each of the two communities and FeNO levels, but only penta-acylated LPS producers appeared to be reduced or absent in individuals with high FeNO. It is likely that the pro-inflammatory effect of hexa-acylated LPS producers is counteracted by the dominance of the more abundant penta-acylated LPS producers in this population-based adult cohort involving mainly healthy individuals.

Oral bacterial composition associated with lung function and lung inflammation in a community-based Norwegian population.

BACKGROUND: The oral cavity is the gateway to the bacteria community in the lung. Disruption of the symbiotic balance of the oral microbiota has been associated with respiratory diseases. However, little is known about the relationship between oral bacteria and respiratory outcomes in the general population. We aimed to describe the associations between oral bacteria, lung function, and lung inflammation in a community-based population. METHODS: Oral (gingival) samples were collected concurrently with spirometry tests in 477 adults (47% males, median age 28 years) from the RHINESSA study in Bergen, Norway. Bacterial DNA from the 16S rRNA gene from gingival fluid were sequenced by Illumina®MiSeq. Lung function was measured using spirometry and measurement of fractional exhaled nitric oxide (FeNO) were performed to examine airway inflammation. Differential abundance analysis was performed using ANCOM-BC, adjusting for weight, education, and smoking. RESULTS: The abundance of the genera Clostridiales, Achromobacter, Moraxella, Flavitalea and Helicobacter were significantly different among those with low FEV1 (< lower limit of normal (LLN)) as compared to normal FEV1 i.e. ≥ LLN. Twenty-three genera differed in abundance between among those with low FVC < LLN as compared to normal FEV1 ≥ LLN. The abundance of 27 genera from phyla Actinobacteria, Bacteroidetes, Firmicutes, Proteobacteria and Sacchribacteria differed significantly between elevated FeNO levels (≥ 50 ppb) compared to FeNO ≤ 25 ppb. CONCLUSION: Oral bacterial composition was significantly different for those with low FEV or FVC as compared to those with normal lung function equal to or higher than LLN. Differential bacterial composition was also observed for elevated FeNO levels.

Relation Between Dietary Protein Intake and Gut Microbiome Composition in Community-Dwelling Older Men: Findings from the Osteoporotic Fractures in Men Study (MrOS).

BACKGROUND: Little is known about the association of specific nutrients, especially proteins, on age-related gut dysbiosis. OBJECTIVES: To determine the associations between the quantity and sources (vegetable and animal) of dietary protein intake and gut microbiome composition in community-dwelling older men. METHODS: We performed a cross-sectional analysis on 775 older men from the Osteoporotic Fractures in Men Study (MrOS) (age 84.2 ± 4.0 y) with available dietary information and stool samples at visit 4 (2014-2016). Protein intake was estimated from a brief FFQ and adjusted to total energy intake. The gut microbiome composition was determined by 16S (v4) sequencing (processed by DADA2 and SILVA). A total of 11,534 amplicon sequence variants (ASVs) were identified and assigned to 21 phyla with dominance of Firmicutes (45%) and Bacteroidetes (43%). We performed α-diversity, ß-diversity, and taxa abundance (by Analysis of Compositions of Microbiomes with Bias Correction [ANCOM-BC]) to determine the associations between protein intake and the gut microbiome. RESULTS: Median protein intake was 0.7 g/(kg body weight · d). Participants with higher energy-adjusted protein intakes had higher Shannon and Chao1 α-diversity indices (P < 0.05). For ß-diversity analysis, participants with higher protein intakes had a different center in weighted and unweighted UniFrac Principal Co-ordinates Analysis (PCoA) compared with those with lower intake (P < 0.05), adjusted for age, race, education, clinical center, batch number, fiber and energy intake, weight, height, and medications. Similarly, higher protein consumptions from either animal or vegetable sources were associated with higher gut microbiome diversity. Several genus-level ASVs, including Christensenellaceae, Veillonella, Haemophilus, and Klebsiella were more abundant in participants with higher protein intakes, whereas Clostridiales bacterium DTU089 and Desulfovibrio were more abundant in participants with lower protein intake (Bonferroni corrected P < 0.05). CONCLUSIONS: We observed significant associations between protein intake and gut microbiome diversity in community-living older men. Further studies are needed to elucidate the mediation role of the gut microbiome on the relation between protein intake and health outcomes in older adults.

Chronic Inhalation Exposure to Antimony Trioxide Exacerbates the MAPK Signaling in Alveolar Bronchiolar Carcinomas in B6C3F1/N Mice.

Antimony trioxide (AT) is used as a flame retardant in fabrics and plastics. Occupational exposure in miners and smelters is mainly through inhalation and dermal contact. Chronic inhalation exposure to AT particulates in B6C3F1/N mice and Wistar Han rats resulted in increased incidences and tumor multiplicities of alveolar/bronchiolar carcinomas (ABCs). In this study, we demonstrated Kras (43%) and Egfr (46%) hotspot mutations in mouse lung tumors (n = 80) and only Egfr (50%) mutations in rat lung tumors (n = 26). Interestingly, there were no differences in the incidences of these mutations in ABCs from rats and mice at exposure concentrations that did and did not exceed the pulmonary overload threshold. There was increased expression of p44/42 mitogen-activated protein kinase (MAPK) (Erk1/2) protein in ABCs harboring mutations in Kras and/or Egfr, confirming the activation of MAPK signaling. Transcriptomic analysis indicated significant alterations in MAPK signaling such as ephrin receptor signaling and signaling by Rho-family GTPases in AT-exposed ABCs. In addition, there was significant overlap between transcriptomic data from mouse ABCs due to AT exposure and human pulmonary adenocarcinoma data. Collectively, these data suggest chronic AT exposure exacerbates MAPK signaling in ABCs and, thus, may be translationally relevant to human lung cancers.

Real-time imaging of asthmatic epithelial cells identifies migratory deficiencies under type-2 conditions.

BACKGROUND: The epithelium is increasingly recognized as a pathologic contributor to asthma and its phenotypes. Although delayed wound closure by asthmatic epithelial cells is consistently observed, underlying mechanisms remain poorly understood, partly due to difficulties in studying dynamic physiologic processes involving polarized multilayered cell systems. Although type-2 immunity has been suggested to play a role, the mechanisms by which repair is diminished are unclear. OBJECTIVES: This study sought to develop and utilize primary multilayered polarized epithelial cell systems, derived from patients with asthma, to evaluate cell migration in response to wounding under type-2 and untreated conditions. METHODS: A novel wounding device for multilayered polarized cells, along with time-lapse live cell/real-time confocal imaging were evaluated under IL-13 and untreated conditions. The influence of inhibition of 15 lipoxygenase (15LO1), a type-2 enzyme, on the process was also addressed. Cell migration patterns were analyzed by high-dimensional frequency modulated Möbius for statistical comparisons. RESULTS: IL-13 stimulation negatively impacts wound healing by altering the total speed, directionality, and acceleration of individual cells. Inhibition 15LO1 partially improved the wound repair through improving total speed. CONCLUSIONS: Migration abnormalities contributed to markedly slower wound closure of IL-13 treated cells, which was modestly reversed by 15LO1 inhibition, suggesting its potential as an asthma therapeutic target. These novel methodologies offer new ways to dynamically study cell movements and identify contributing pathologic processes.

Temporal and Spatial Changes in the Microbiome Following Pediatric Severe Traumatic Brain Injury.

OBJECTIVES: The microbiome may be affected by trauma and critical illness. Many studies of the microbiome in critical illness are restricted to a single body site or time point and confounded by preexisting conditions. We report temporal and spatial alterations in the microbiome of previously healthy children with severe traumatic brain injury (TBI). DESIGN: We collected oral, rectal, and skin swabs within 72 hours of admission and then twice weekly until ICU discharge. Samples were analyzed by 16S rRNA gene amplicon sequencing. Children undergoing elective outpatient surgery served as controls. Alpha and beta diversity comparisons were performed with Phyloseq, and differentially abundant taxa were predicted using Analysis of Composition of Microbiomes. SETTING: Five quaternary-care PICUs. PATIENTS: Patients less than 18 years with severe TBI requiring placement of an intracranial pressure monitor. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Three hundred twenty-seven samples were analyzed from 23 children with severe TBI and 35 controls. The community composition of initial oral (F = 3.2756, R2 = 0.0535, p = 0.012) and rectal (F = 3.0702, R2 = 0.0649, p = 0.007) samples differed between TBI and control patients. Rectal samples were depleted of commensal bacteria from Ruminococcaceae, Bacteroidaceae, and Lachnospiraceae families and enriched in Staphylococcaceae after TBI (p < 0.05). In exploratory analyses, antibiotic exposure, presence of an endotracheal tube, and occurrence of an infection were associated with greater differences of the rectal and oral microbiomes between TBI patients and healthy controls, whereas enteral nutrition was associated with smaller differences (p < 0.05). CONCLUSIONS: The microbiome of children with severe TBI is characterized by early depletion of commensal bacteria, loss of site specificity, and an enrichment of potential pathogens. Additional studies are needed to determine the impact of these changes on clinical outcomes.

House dust microbiota in relation to adult asthma and atopy in a US farming population.

BACKGROUND: Bacterial exposure from house dust has been associated with asthma and atopy in children but whether these relationships are present in adults remains unclear. OBJECTIVE: We sought to examine associations of house dust microbiota with adult asthma, atopy, and hay fever. METHODS: Vacuumed bedroom dust samples from the homes of 879 participants (average age, 62 years) in the Agricultural Lung Health Study, a case-control study of asthma nested within a farming cohort, were subjected to 16S rRNA amplicon sequencing to characterize bacterial communities. We defined current asthma and hay fever using questionnaires and current atopy by blood specific IgE level > 0.70 IU/mL to 1 or more of 10 common allergens. We used linear regression to examine whether overall within-sample bacterial diversity differed by outcome, microbiome regression-based kernel association test to evaluate whether between-sample bacterial community compositions differed by outcome, and analysis of composition of microbiomes to identify differentially abundant bacterial taxa. RESULTS: Overall diversity of bacterial communities in house dust was similar by asthma status but was lower (P < .05) with atopy or hay fever. Many individual bacterial taxa were differentially abundant (false-discovery rate, <0.05) by asthma, atopy, or hay fever. Several taxa from Cyanobacteria, Bacteroidetes, and Fusobacteria were more abundant with asthma, atopy, or hay fever. In contrast, several taxa from Firmicutes were more abundant in homes of individuals with adequately controlled asthma (vs inadequately controlled asthma), individuals without atopy, or individuals without hay fever. CONCLUSIONS: Microbial composition of house dust may influence allergic outcomes in adults.

Order restricted inference in chronobiology.

This paper is motivated by applications in oscillatory systems where researchers are typically interested in discovering components of those systems that display rhythmic temporal patterns. The contributions of this paper are twofold. First, a methodology is developed based on a circular signal plus error model that is defined using order restrictions. This mathematical formulation of rhythmicity is simple, easily interpretable and very flexible, with the latter property derived from the nonparametric formulation of the signal. Second, we address various commonly encountered problems in the analysis of oscillatory systems data. Specifically, we propose a methodology for (a) detecting rhythmic signals in an oscillatory system and (b) estimating the unknown sampling time that occurs when tissues are obtained from subjects whose time of death is unknown. The proposed methodology is computationally efficient, outperforms the existing methods, and is broadly applicable to address a wide range of questions related to oscillatory systems.

Unique microRNA alterations in hepatocellular carcinomas arising either spontaneously or due to chronic exposure to Ginkgo biloba extract (GBE) in B6C3F1/N mice.

Ginkgo biloba extract (GBE) is used in traditional Chinese medicine as a herbal supplement for improving memory. Exposure of B6C3F1/N mice to GBE in a 2-year National Toxicology Program (NTP) bioassay resulted in a dose-dependent increase in hepatocellular carcinomas (HCC). To identify key microRNAs that modulate GBE-induced hepatocarcinogenesis, we compared the global miRNA expression profiles in GBE-exposed HCC (GBE-HCC) and spontaneous HCC (SPNT-HCC) with age-matched vehicle control normal livers (CNTL) from B6C3F1/N mice. The number of differentially altered miRNAs in GBE-HCC and SPNT-HCC was 74 (52 up and 22 down) and 33 (15 up and 18 down), respectively. Among the uniquely differentially altered miRNAs in GBE-HCC, miR-31 and one of its predicted targets, Cdk1 were selected for functional validation. A potential miRNA response element (MRE) in the 3'-untranslated regions (3'-UTR) of Cdk1 mRNA was revealed by in silico analysis and confirmed by luciferase assays. In mouse hepatoma cell line HEPA-1 cells, we demonstrated an inverse correlation between miR-31 and CDK1 protein levels, but no change in Cdk1 mRNA levels, suggesting a post-transcriptional effect. Additionally, a set of miRNAs (miRs-411, 300, 127, 134, 409-3p, and 433-3p) that were altered in the GBE-HCCs were also altered in non-tumor liver samples from the 90-day GBE-exposed group compared to the vehicle control group, suggesting that some of these miRNAs could serve as potential biomarkers for GBE exposure or hepatocellular carcinogenesis. These data increase our understanding of miRNA-mediated epigenetic regulation of GBE-mediated hepatocellular carcinogenesis in B6C3F1/N mice.

Psychosocial and behavioral pathways of metabolic syndrome in cancer caregivers.

OBJECTIVE: Cancer caregivers are at increased risk for cardiovascular disease (CVD) and mortality. The aims of this study were to examine psychosocial and behavioral predictors of metabolic syndrome, an intermediate endpoint of CVD. METHODS: Cancer caregivers were administered a battery of questionnaires assessing sociodemographic characteristics, depressive symptoms, perceived stress, caregiver quality of life, sleep, physical activity, alcohol and tobacco use, social support, relationship quality, and loneliness. Metabolic syndrome was defined using the American Heart Association guidelines and the National Cholesterol Education Program's Adult Treatment Panel (ATP) III, which includes the presence of at least three of the following abnormalities: blood pressure, glucose, abdominal girth, high-density lipoprotein (HDL), and triglycerides. RESULTS: Of the 104 caregivers, 77% were female, 94% were Caucasian, and the mean age was 59.5 (SD = 12.8). Of the 104 caregivers, 35.6% reported depressive symptoms in the clinical range of the Center for Epidemiologic Studies-Depression (CES-D) and 69.2% reported Perceived Stress Scale scores at least one standard deviation above the general population norms. A total of 16.3% of caregivers currently used tobacco, 28.8% consumed alcohol, and 26.7% were overweight (BMI = 25-29.9) and 48.5% were obese (BMI ≥ 30). Forty-nine percent of the caregivers met the criteria for metabolic syndrome. After age, gender, and race were adjusted, the following remained as significant predictors of metabolic syndrome: low levels of caregiver quality of life (Odds Ratio (OR) = 1.067; 95% CI, 1.019-1.117; P = .006), high levels of hostility (OR = 1.142; 95% CI, 1.030-1.267; P = .012), and current alcohol use (OR = 4.193; 95% CI, 1.174-14.978; P = .027). CONCLUSION: Development of interventions to reduce the risk of metabolic syndrome in cancer caregivers is recommended.

Immunization with a heat-killed preparation of the environmental bacterium Mycobacterium vaccae promotes stress resilience in mice.

The prevalence of inflammatory diseases is increasing in modern urban societies. Inflammation increases risk of stress-related pathology; consequently, immunoregulatory or antiinflammatory approaches may protect against negative stress-related outcomes. We show that stress disrupts the homeostatic relationship between the microbiota and the host, resulting in exaggerated inflammation. Repeated immunization with a heat-killed preparation of Mycobacterium vaccae, an immunoregulatory environmental microorganism, reduced subordinate, flight, and avoiding behavioral responses to a dominant aggressor in a murine model of chronic psychosocial stress when tested 1-2 wk following the final immunization. Furthermore, immunization with M. vaccae prevented stress-induced spontaneous colitis and, in stressed mice, induced anxiolytic or fear-reducing effects as measured on the elevated plus-maze, despite stress-induced gut microbiota changes characteristic of gut infection and colitis. Immunization with M. vaccae also prevented stress-induced aggravation of colitis in a model of inflammatory bowel disease. Depletion of regulatory T cells negated protective effects of immunization with M. vaccae on stress-induced colitis and anxiety-like or fear behaviors. These data provide a framework for developing microbiome- and immunoregulation-based strategies for prevention of stress-related pathologies.

Threshold Knot Selection for Large-Scale Spatial Models With Applications to the Deepwater Horizon Disaster.

Large spatial datasets are typically modeled through a small set of knot locations; often these locations are specified by the investigator by arbitrary criteria. Existing methods of estimating the locations of knots assume their number is known a priori, or are otherwise computationally intensive. We develop a computationally efficient method of estimating both the location and number of knots for spatial mixed effects models. Our proposed algorithm, Threshold Knot Selection (TKS), estimates knot locations by identifying clusters of large residuals and placing a knot in the centroid of those clusters. We conduct a simulation study showing TKS in relation to several comparable methods of estimating knot locations. Our case study utilizes data of particulate matter concentrations collected during the course of the response and clean-up effort from the 2010 Deepwater Horizon oil spill in the Gulf of Mexico.

Structural zeros in high-dimensional data with applications to microbiome studies.

This paper is motivated by the recent interest in the analysis of high-dimensional microbiome data. A key feature of these data is the presence of "structural zeros" which are microbes missing from an observation vector due to an underlying biological process and not due to error in measurement. Typical notions of missingness are unable to model these structural zeros. We define a general framework which allows for structural zeros in the model and propose methods of estimating sparse high-dimensional covariance and precision matrices under this setup. We establish error bounds in the spectral and Frobenius norms for the proposed estimators and empirically verify them with a simulation study. The proposed methodology is illustrated by applying it to the global gut microbiome data of Yatsunenko and others (2012. Human gut microbiome viewed across age and geography. Nature 486, 222-227). Using our methodology we classify subjects according to the geographical location on the basis of their gut microbiome.

Reduced Disc Shedding and Phagocytosis of Photoreceptor Outer Segment Contributes to Kava Kava Extract-induced Retinal Degeneration in F344/N Rats.

There was a significant increase in the incidence of retinal degeneration in F344/N rats chronically exposed to Kava kava extract (KKE) in National Toxicology Program (NTP) bioassay. A retrospective evaluation of these rat retinas indicated a similar spatial and morphological alteration as seen in light-induced retinal degeneration in albino rats. Therefore, it was hypothesized that KKE has a potential to exacerbate the light-induced retinal degeneration. To investigate the early mechanism of retinal degeneration, we conducted a 90-day F344/N rat KKE gavage study at doses of 0 and 1.0 g/kg (dose which induced retinal degeneration in the 2-year NTP rat KKE bioassay). The morphological evaluation indicated reduced number of phagosomes in the retinal pigment epithelium (RPE) of the superior retina. Transcriptomic alterations related to retinal epithelial homeostasis and melatoninergic signaling were observed in microarray analysis. Phagocytosis of photoreceptor outer segment by the underlying RPE is essential to maintain the homeostasis of the photoreceptor layer and is regulated by melatonin signaling. Therefore, reduced photoreceptor outer segment disc shedding and subsequent lower number of phagosomes in the RPE and alterations in the melatonin pathway may have contributed to the increased incidences of retinal degeneration observed in F344/N rats in the 2-year KKE bioassay.

Order restricted inference for oscillatory systems for detecting rhythmic signals.

MOTIVATION: Many biological processes, such as cell cycle, circadian clock, menstrual cycles, are governed by oscillatory systems consisting of numerous components that exhibit rhythmic patterns over time. It is not always easy to identify such rhythmic components. For example, it is a challenging problem to identify circadian genes in a given tissue using time-course gene expression data. There is a great potential for misclassifying non-rhythmic as rhythmic genes and vice versa. This has been a problem of considerable interest in recent years. In this article we develop a constrained inference based methodology called Order Restricted Inference for Oscillatory Systems (ORIOS) to detect rhythmic signals. Instead of using mathematical functions (e.g. sinusoidal) to describe shape of rhythmic signals, ORIOS uses mathematical inequalities. Consequently, it is robust and not limited by the biologist's choice of the mathematical model. We studied the performance of ORIOS using simulated as well as real data obtained from mouse liver, pituitary gland and data from NIH3T3, U2OS cell lines. Our results suggest that, for a broad collection of patterns of gene expression, ORIOS has substantially higher power to detect true rhythmic genes in comparison to some popular methods, while also declaring substantially fewer non-rhythmic genes as rhythmic. AVAILABILITY AND IMPLEMENTATION: A user friendly code implemented in R language can be downloaded from http://www.niehs.nih.gov/research/atniehs/labs/bb/staff/peddada/index.cfm CONTACT: peddada@niehs.nih.gov.

Parkinson's disease and Parkinson's disease medications have distinct signatures of the gut microbiome.

BACKGROUND: There is mounting evidence for a connection between the gut and Parkinson's disease (PD). Dysbiosis of gut microbiota could explain several features of PD. OBJECTIVE: The objective of this study was to determine if PD involves dysbiosis of gut microbiome, disentangle effects of confounders, and identify candidate taxa and functional pathways to guide research. METHODS: A total of 197 PD cases and 130 controls were studied. Microbial composition was determined by 16S rRNA gene sequencing of DNA extracted from stool. Metadata were collected on 39 potential confounders including medications, diet, gastrointestinal symptoms, and demographics. Statistical analyses were conducted while controlling for potential confounders and correcting for multiple testing. We tested differences in the overall microbial composition, taxa abundance, and functional pathways. RESULTS: Independent microbial signatures were detected for PD (P = 4E-5), participants' region of residence within the United States (P = 3E-3), age (P = 0.03), sex (P = 1E-3), and dietary fruits/vegetables (P = 0.01). Among patients, independent signals were detected for catechol-O-methyltransferase-inhibitors (P = 4E-4), anticholinergics (P = 5E-3), and possibly carbidopa/levodopa (P = 0.05). We found significantly altered abundances of the Bifidobacteriaceae, Christensenellaceae, [Tissierellaceae], Lachnospiraceae, Lactobacillaceae, Pasteurellaceae, and Verrucomicrobiaceae families. Functional predictions revealed changes in numerous pathways, including the metabolism of plant-derived compounds and xenobiotics degradation. CONCLUSION: PD is accompanied by dysbiosis of gut microbiome. Results coalesce divergent findings of prior studies, reveal altered abundance of several taxa, nominate functional pathways, and demonstrate independent effects of PD medications on the microbiome. The findings provide new leads and testable hypotheses on the pathophysiology and treatment of PD. © 2017 International Parkinson and Movement Disorder Society.

A multiple testing procedure for multi-dimensional pairwise comparisons with application to gene expression studies.

BACKGROUND: Often researchers are interested in comparing multiple experimental groups (e.g. tumor size) with a reference group (e.g. normal tissue) on the basis of thousands of features (e.g. genes) and determine if a differentially expressed feature is up or down regulated in a pairwise comparison. There are two sources of false discoveries, one due to multiple testing involving several pairwise comparisons and the second due to falsely declaring a feature to be up (or down) regulated when it is not (known as directional error). Together, the total error rate is called the mixed directional false discovery rate (mdFDR). RESULTS: We develop a general powerful mdFDR controlling testing procedure and illustrate the methodology by analyzing uterine fibroid gene expression data (PLoS ONE 8:63909, 2013). We identify several differentially expressed genes (DEGs) and pathways that are specifically enriched according to the size of a uterine fibroid. CONCLUSIONS: The proposed general procedure strongly controls mdFDR. Several specific methodologies can be derived from this general methodology by using appropriate testing procedures at different steps of the general procedure. Thus we are providing a general framework for making multiple pairwise comparisons. Our analysis of the uterine fibroid growth gene expression data suggests that molecular characteristics of a fibroid changes with size. Our powerful methodology allowed us to draw several interesting conclusions regarding the molecular characteristics of uterine fibroids. For example, IL-1 signaling pathway (Sci STKE 2003:3, 2003), associated with inflammation and known to activate prostaglandins that are implicated in the progression of fibroids, is significantly enriched only in small tumors (volume < 5.7 cm (3)). It appears that the molecular apparatus necessary for fibroid growth and development is established during tumor development. A complete list of all DEGs and the corresponding enriched pathways according to tumor size is provided for researchers to mine these data. Identification of these DEGs and the pathways may potentially have clinical implications.