Browse
Submit Data
Databases
API
Help

Dataset Information

0 Views

0 Connections

0 Citations

0 Reanalyses

0 Downloads

Omics score: 0

RFtest: A Robust and Flexible Community-Level Test for Microbiome Data Powerfully Detects Phylogenetically Clustered Signals.

ABSTRACT: Random forest is considered as one of the most successful machine learning algorithms, which has been widely used to construct microbiome-based predictive models. However, its use as a statistical testing method has not been explored. In this study, we propose "Random Forest Test" (RFtest), a global (community-level) test based on random forest for high-dimensional and phylogenetically structured microbiome data. RFtest is a permutation test using the generalization error of random forest as the test statistic. Our simulations demonstrate that RFtest has controlled type I error rates, that its power is superior to competing methods for phylogenetically clustered signals, and that it is robust to outliers and adaptive to interaction effects and non-linear associations. Finally, we apply RFtest to two real microbiome datasets to ascertain whether microbial communities are associated or not with the outcome variables.

SUBMITTER: Zhang L

PROVIDER: S-EPMC8819960 | biostudies-literature | 2021

REPOSITORIES: biostudies-literature

ACCESS DATA

Json Xml

Publications

RFtest: A Robust and Flexible Community-Level Test for Microbiome Data Powerfully Detects Phylogenetically Clustered Signals.

Zhang Lujun L Wang Yanshan Y Chen Jingwen J Chen Jun J

Frontiers in genetics 20220124

Random forest is considered as one of the most successful machine learning algorithms, which has been widely used to construct microbiome-based predictive models. However, its use as a statistical testing method has not been explored. In this study, we propose "Random Forest Test" (RFtest), a global (community-level) test based on random forest for high-dimensional and phylogenetically structured microbiome data. RFtest is a permutation test using the generalization error of random forest as the ...[more]

PMID: 35140735

Similar Datasets

Arbuscular mycorrhizal fungal communities are phylogenetically clustered at small scales.

Project description:Next-generation sequencing technologies with markers covering the full Glomeromycota phylum were used to uncover phylogenetic community structure of arbuscular mycorrhizal fungi (AMF) associated with Festuca brevipila. The study system was a semi-arid grassland with high plant diversity and a steep environmental gradient in pH, C, N, P and soil water content. The AMF community in roots and rhizosphere soil were analyzed separately and consisted of 74 distinct operational taxonomic units (OTUs) in total. Community-level variance partitioning showed that the role of environmental factors in determining AM species composition was marginal when controlling for spatial autocorrelation at multiple scales. Instead, phylogenetic distance and spatial distance were major correlates of AMF communities: OTUs that were more closely related (and which therefore may have similar traits) were more likely to co-occur. This pattern was insensitive to phylogenetic sampling breadth. Given the minor effects of the environment, we propose that at small scales closely related AMF positively associate through biotic factors such as plant-AMF filtering and interactions within the soil biota.

| S-EPMC4992081 | biostudies-literature

Phylogenetically clustered extinction risks do not substantially prune the Tree of Life.

Project description:Anthropogenic activities have increased the rate of biological extinction many-fold. Recent empirical studies suggest that projected extinction may lead to extensive loss to the Tree of Life, much more than if extinction were random. One suggested cause is that extinction risk is heritable (phylogenetically patterned), such that entire higher groups will be lost. We show here with simulation that phylogenetically clustered extinction risks are necessary but not sufficient for the extensive loss of phylogenetic diversity (PD) compared to random extinction. We simulated Yule trees and evolved extinction risks at various levels of heritability (measured using Pagel's ?). At most levels of heritability (? in range of 0 to 10), mean values of extinction risk (range 0.25 to 0.75), tree sizes (64 to 128 tips), tree balance and temporal heterogeneity of diversification rates (Yule and coalescent trees), extinction risks do not substantially increase the loss of PD in these trees when compared to random extinction. The maximum loss of PD (20% above random) was only associated with the combination of extremely excessive values of phylogenetic signal, high mean species' extinction probabilities, and extreme (coalescent) tree shapes. Interestingly, we also observed a decline in the rate of increase in the loss of PD at high phylogenetic clustering (? ? 10) of extinction risks. Our results suggest that the interplay between various aspects of tree shape and a predisposition of higher extinction risks in species-poor clades is required to explain the substantial pruning of the Tree of Life.

| S-EPMC3154466 | biostudies-other

Robust phylogenetic tree-based microbiome association test using repeatedly measured data for composition bias.

Project description:BackgroundThe effects of microbiota on the host phenotypes can differ substantially depending on their age. Longitudinally measured microbiome data allow for the detection of the age modification effect and are useful for the detection of microorganisms related to the progression of disease whose identification change over time. Moreover, longitudinal analysis facilitates the estimation of the within-subject covariate effect, is robust to the between-subject confounders, and provides better evidence for the causal relationship than cross-sectional studies. However, this method of analysis is limited by compositional bias, and few statistical methods can estimate the effect of microbiota on host diseases with repeatedly measured 16S rRNA gene data. Herein, we propose mTMAT, which is applicable to longitudinal microbiome data and is robust to compositional bias.ResultsmTMAT normalized the microbial abundance and utilized the ratio of the pooled abundance for association analysis. mTMAT is based on generalized estimating equations with a robust variance estimator and can be applied to repeatedly measured microbiome data. The robustness of mTMAT against compositional bias is underscored by its utilization of abundance ratios.ConclusionsWith extensive simulation studies, we showed that mTMAT is statistically relatively powerful and is robust to compositional bias. mTMAT enables detection of microbial taxa associated with host diseases using repeatedly measured 16S rRNA gene data and can provide deeper insights into bacterial pathology.

| S-EPMC11887327 | biostudies-literature

Chromosomal regions specific to pathogenic isolates of Escherichia coli have a phylogenetically clustered distribution.

Project description:We studied the ancestry of virulence-associated genes in Escherichia coli by examining chromosomal regions specific to pathogenic isolates. The four virulence determinants examined were the alpha-hemolysin (hly) loci hlyI and hlyII, the type II capsule gene cluster kps, and the P (pap) and S (sfa) fimbria gene clusters. All four loci were shown previously to be associated with pathogenicity islands of uropathogenic E. coli isolates. The hly, kps, sfa, and pap regions each have an unexpected clustered distribution among the E. coli collection of reference (ECOR) strains, but all these regions were absent from a collection of diarrheagenic E. coli isolates. Strains in the ECOR subgroup B2 typically had a combination of at least three of the four loci, and all strains in subgroup D had a copy of the kps and pap clusters. In contrast, only four strains in subgroup A had either hly, kps, sfa, or pap, and no subgroup A strains had all four together. Strains of subgroup B1 were devoid of all four virulence regions, with the exception of one isolate that had a copy of the sfa gene cluster. This phylogenetic distribution of strain-specific sequences corresponds to the ECOR groups with the largest genome size, namely, B2 and D. We propose that the pathogenicity islands are ancestral to subgroups B2 and D and were acquired after speciation, with subsequent horizontal transfer into some group A, B1, and E lineages. These results suggest that the hly, kps, sfa, and pap pathogenicity determinants may play a role in the evolution of enteric bacteria quite apart from, and perhaps with precedence over, their ability to cause disease.

| S-EPMC107003 | biostudies-literature

Inhospitable sweetness: nectar filtering of pollinator-borne inocula leads to impoverished, phylogenetically clustered yeast communities.

Project description:Identifying the rules and mechanisms that determine the composition and diversity of naturally co-occurring species assemblages is a central topic in community ecology. Although micro-organisms represent the 'unseen majority' of species, individuals and biomass in many ecosystems and play pivotal roles in community development and function, the study of the factors influencing the assembly of microbial communities has lagged behind that of plant and animal communities. In this paper, we investigate experimentally the mechanisms accounting for the low species richness of yeast communities inhabiting the nectar of the bumble-bee-pollinated Helleborus foetidus (Ranunculaceae), and explore the relationships between community assembly rules and phylogenetic relatedness. By comparing yeast communities on the glossae of foraging bumble-bees (the potential species pool) with those eventually establishing in virgin nectar probed with bee glossae (the realized community), we address the questions: (i) does nectar filter yeast inocula, so that the communities eventually established there are not random subsamples of species on bumble-bee glossae? and (ii) do yeast communities establishing in H. foetidus nectar exhibit some phylogenetic bias relative to the species pool on bumble-bee glossae? Results show that nectar filtering leads to species-poor, phylogenetically clustered yeast communities that are a predictable subset of pollinator-borne inocula. Such strong habitat filtering is probably due to H. foetidus nectar representing a harsh environment for most yeasts, where only a few phylogenetically related nectar specialists physiologically endowed to tolerate a combination of high osmotic pressure and fungicidal compounds are able to develop.

| S-EPMC2842741 | biostudies-literature

A fast small-sample kernel independence test for microbiome community-level association analysis.

Project description:To fully understand the role of microbiome in human health and diseases, researchers are increasingly interested in assessing the relationship between microbiome composition and host genomic data. The dimensionality of the data as well as complex relationships between microbiota and host genomics pose considerable challenges for analysis. In this article, we apply a kernel RV coefficient (KRV) test to evaluate the overall association between host gene expression and microbiome composition. The KRV statistic can capture nonlinear correlations and complex relationships among the individual data types and between gene expression and microbiome composition through measuring general dependency. Testing proceeds via a similar route as existing tests of the generalized RV coefficients and allows for rapid p-value calculation. Strategies to allow adjustment for confounding effects, which is crucial for avoiding misleading results, and to alleviate the problem of selecting the most favorable kernel are considered. Simulation studies show that KRV is useful in testing statistical independence with finite samples given the kernels are appropriately chosen, and can powerfully identify existing associations between microbiome composition and host genomic data while protecting type I error. We apply the KRV to a microbiome study examining the relationship between host transcriptome and microbiome composition within the context of inflammatory bowel disease and are able to derive new biological insights and provide formal inference on prior qualitative observations.

| S-EPMC5592124 | biostudies-literature

Colocalization of GWAS and eQTL Signals Detects Target Genes.

Project description:The vast majority of genome-wide association study (GWAS) risk loci fall in non-coding regions of the genome. One possible hypothesis is that these GWAS risk loci alter the individual's disease risk through their effect on gene expression in different tissues. In order to understand the mechanisms driving a GWAS risk locus, it is helpful to determine which gene is affected in specific tissue types. For example, the relevant gene and tissue could play a role in the disease mechanism if the same variant responsible for a GWAS locus also affects gene expression. Identifying whether or not the same variant is causal in both GWASs and expression quantitative trail locus (eQTL) studies is challenging because of the uncertainty induced by linkage disequilibrium and the fact that some loci harbor multiple causal variants. However, current methods that address this problem assume that each locus contains a single causal variant. In this paper, we present eCAVIAR, a probabilistic method that has several key advantages over existing methods. First, our method can account for more than one causal variant in any given locus. Second, it can leverage summary statistics without accessing the individual genotype data. We use both simulated and real datasets to demonstrate the utility of our method. Using publicly available eQTL data on 45 different tissues, we demonstrate that eCAVIAR can prioritize likely relevant tissues and target genes for a set of glucose- and insulin-related trait loci.

| S-EPMC5142122 | biostudies-literature

Robust identification of noncoding RNA from transcriptomes requires phylogenetically-informed sampling.

Project description:Noncoding RNAs are integral to a wide range of biological processes, including translation, gene regulation, host-pathogen interactions and environmental sensing. While genomics is now a mature field, our capacity to identify noncoding RNA elements in bacterial and archaeal genomes is hampered by the difficulty of de novo identification. The emergence of new technologies for characterizing transcriptome outputs, notably RNA-seq, are improving noncoding RNA identification and expression quantification. However, a major challenge is to robustly distinguish functional outputs from transcriptional noise. To establish whether annotation of existing transcriptome data has effectively captured all functional outputs, we analysed over 400 publicly available RNA-seq datasets spanning 37 different Archaea and Bacteria. Using comparative tools, we identify close to a thousand highly-expressed candidate noncoding RNAs. However, our analyses reveal that capacity to identify noncoding RNA outputs is strongly dependent on phylogenetic sampling. Surprisingly, and in stark contrast to protein-coding genes, the phylogenetic window for effective use of comparative methods is perversely narrow: aggregating public datasets only produced one phylogenetic cluster where these tools could be used to robustly separate unannotated noncoding RNAs from a null hypothesis of transcriptional noise. Our results show that for the full potential of transcriptomics data to be realized, a change in experimental design is paramount: effective transcriptomics requires phylogeny-aware sampling.

| S-EPMC4214555 | biostudies-literature

Rainfall-induced carbon dioxide pulses result from sequential resuscitation of phylogenetically clustered microbial groups.

Project description:The pulse of carbon dioxide (CO(2)) resulting from the first rainfall after the dry summer in Mediterranean ecosystems is so large that it is well documented at the landscape scale, with the CO(2) released in a few days comparable in magnitude to the annual net carbon exchange of many terrestrial ecosystems. Although the origin of this CO(2) is debated, we show that the pulse of CO(2) is produced by a three-step resuscitation of the indigenous microbial community. Specific phylogenetic groups of microorganisms activate and contribute to the CO(2) pulse at different times after a simulation of the first rainfall following the severe summer drought. Differential resuscitation was evident within 1 h of wet-up, with three primary response strategies apparent according to patterns of relative ribosomal quantity. Most bacteria could be classified as rapid responders (within 1 h of wet-up), intermediate responders (between 3 and 24 h after wet-up), or delayed responders (24-72 h after wet-up). Relative ribosomal quantities of rapid responders were as high in the prewet dry soils as at any other time, suggesting that specific groups of organisms may be poised to respond to the wet-up event, in that they preserve their capacity to synthesize proteins rapidly. Microbial response patterns displayed phylogenetic clustering and were primarily conserved at the subphylum level, suggesting that resuscitation strategies after wet-up of dry soil may be a phylogenetically conserved ecological trait.

| S-EPMC3390866 | biostudies-other

Flexible propensity score estimation strategies for clustered data in observational studies.

Project description:Existing studies have suggested superior performance of nonparametric machine learning over logistic regression for propensity score estimation. However, it is unclear whether the advantages of nonparametric propensity score modeling are carried to settings where there is clustering of individuals, especially when there is unmeasured cluster-level confounding. In this work we examined the performance of logistic regression (all main effects), Bayesian additive regression trees and generalized boosted modeling for propensity score weighting in clustered settings, with the clustering being accounted for by including either cluster indicators or random intercepts. We simulated data for three hypothetical observational studies of varying sample and cluster sizes. Confounders were generated at both levels, including a cluster-level confounder that is unobserved in the analyses. A binary treatment and a continuous outcome were generated based on seven scenarios with varying relationships between the treatment and confounders (linear and additive, nonlinear/nonadditive, nonadditive with the unobserved cluster-level confounder). Results suggest that when the sample and cluster sizes are large, nonparametric propensity score estimation may provide better covariate balance, bias reduction, and 95% confidence interval coverage, regardless of the degree of nonlinearity or nonadditivity in the true propensity score model. When the sample or cluster sizes are small, however, nonparametric approaches may become more vulnerable to unmeasured cluster-level confounding and thus may not be a better alternative to multilevel logistic regression. We applied the methods to the National Longitudinal Study of Adolescent to Adult Health data, estimating the effect of team sports participation during adolescence on adulthood depressive symptoms.

| S-EPMC9996644 | biostudies-literature

OmicsDI is part of the ELIXIR infrastructure

OmicsDI is an Elixir interoperability service. Learn more ›

Tweets

OmicsDI Databases

PRIDE
PeptideAtlas
MassIVE
JPOST Repository
Physiome Model Repository

EGA
EVA
ENA
LINCS
PAXDB
Cell Collective

MetaboLights
Metabolomics Workbench
MetabolomeExpress
GNPS
BioModels
FAIRDOMHub

ArrayExpress
dbGaP
ExpressionAtlas
GEO
NODE

Information

Databases
Help
API
Contact us
Code on GitHub
Terms of use
Submit Data