AOV1way: One-Way ANOVA

View source: R/statistics_correlations_functions.R View source: R/Package-Code5-Compiled.R

AOV1wayR Documentation

One-Way ANOVA

Description

A one-way ANOVA for the samples in the supplied data matrix based on annotations stored in the params list object.

Usage

AOV1way(data.to.aov, category, pthreshold = 0.05, additional.report = "NONE")

Arguments

data.to.aov

numeric data matrix with samples/observations in the columns and genes/variables in the rows

category

Character string pointing to a column of the annotations data frame stores in params$annotations

pthreshold

pvalue threshold for which genes will be considered significant

additional.report

Option to include full output of aov or TukeyHSD, options are "NONE", "AOV","TUKEY", or "ALL"

Details

Runs a standard one-way ANOVA and subsequent TukeyHSD on the data provided using stored annotations as the categories for ANOVA groupings

Value

~List object

AOV.Results

dataframe giving the F statistic and pvalue of each gene in the original input matrix

Sig.Genes

character vector of the genes with pvalues below the specified pthreshold

NonSig.Genes

character vector of the genes with pvalues that did not pass the specified pthreshold

Tukey.pVals

dataframe containing the adjusted pvalues from TukeyHSD for each contrast for each gene in the original input matrix

Tukey.diffs

dataframe containing the difference values from TukeyHSD for each contrast for each gene in the original input matrix

optional items

AOV.output

list object containing the aov output for all genes in the original input matrix

Tukey.output

list object containing the TUkeyHSD output for all genes in the original input matrix

Note

The function itself determines significant and non-significant genes based on the pthreshold provided. However, the AOV.Results provdied in the output supply the F value and p value, additional information can also be retrieved by setting additional report = TRUE, user can use any of this information for custom filtering of results.

For cases where there are no non-missing arguments for one or more of the groups in the supplied categories, a warning will be generated indicating which genes or observations this is true for. If there is only one group with no non-missing arguments, that gene or observation will be removed from the analysis (and will be indicated with a warning). Please pay attention to these warnings as they may affect how the results should be interpretted.

Author(s)

~~Alison Moss~~

Examples

##initiate parameteres and set up annotations

initiate_params()
data(RAGP_norm)
data(RAGP_annots)
set_annotations(RAGP_annots)

##AOV for Connectivity
state.aov <- AOV1way(RAGP_norm, "State")
head(state.aov$AOV.Results)
head(state.aov$Sig.Genes)

##visualize significant genes in heatmap
set_annot_samps("State")
myHeatmapByAnnotation(RAGP_norm, state.aov$Sig.Genes, groupings = "State")


##AOV for Animal
animal.aov <- AOV1way(RAGP_norm, "Animal", pthreshold = 0.01)
summary(animal.aov)
head(animal.aov$Tukey.pVals) ##look at pvalues from tukey post hoc analysis


axm323/dataVisEasy documentation built on Feb. 1, 2024, 11:53 p.m.