looPA: Microbiome feature identification using leave-one-out with...

In YushuShi/looPA: Microbiome Feature Identification Using Leave-One-Out with Permanova Assessment

Description

looPA is a permutation based method, which can account for phylogenetic relatedness between taxonomic features and identify important features for further investigation.

Usage

looPA(otutable,taxonomy, sampleInfo, outcomeVar, numRep=200,useMoreCores=TRUE,
tree=NULL,distanceMetric="Bray Curtis")

Arguments

otutable	OTU table. Rows correspond to OTUs while columns correspond to samples. Row names (OTU names) and column names(sample names) must be provided.
taxonomy	Taxonomy table. Rows correspond to OTUs while columns correspond to the taxonomic levels. Row names (OTU names) must match the row names of the OTU table.
sampleInfo	Patient information. Rows correspond to samples while columns correspond to covariates. Row names (sample names) must match the column names of the OTU table.
outcomeVar	The outcome of interest. For now we only accept single covariate.
numRep	Number of repeated PERMANOVA test for each feature. Default is 200.
useMoreCores	Shall we use more cores of the computer for this job? If you select TRUE (recommended), the number of cores used will be the number of cores of the computer -1. If you give an integer value, looPA will use it as the number of cores you want to use.
tree	The phylogenetic information correspond to OTU table.
distanceMetric	The distance metric used in PERMANOVA test. We offer three choices, "Bray Curtis" (the default), "Unweighted UniFrac"(requires tree information), and "Weighted UniFrac"(requires tree information).

Value

looPAresult	A dataframe with the names of the selected taxa and their medians and confidence intervals over repeated PERMANOVA tests.
looPAplot	A plot showing the selected features.

Examples

library(vegan)
library(GUniFrac)
library(ape)
library(ggplot2)
library(foreach)
library(parallel)
library(iterators)
library(doParallel)
library(looPA)
options(stringsAsFactors = FALSE)

outcomeVar<-"Response"

looPAResult<-looPA(otutable,taxonomy, sampleInfo, outcomeVar,
tree=tree,distanceMetric="Weighted UniFrac")

YushuShi/looPA documentation built on May 7, 2020, 9:38 p.m.