calculate_anova {MetAlyzer}

One-way ANOVA

Description

This method performs a one-way ANOVA on the grouped aggregated_data (the categorical variable is removed from grouping first). The vector of the categorical variable needs to have at least two levels after removing NAs from the dependent variable vector. Otherwise a vector of NA is returned. A Tukey post-hoc test is then used to determine group names, starting with "A" followed by further letters. These group names are added to aggregated_data in the column ANOVA_Group. Thereby, metabolites can be identified which are significantly higher in one or more of the categorical variable compared to all other for each metabolite.

Usage

calculate_anova(
  metalyzer_se,
  categorical,
  groups = NULL,
  impute_perc_of_min = 0.2,
  impute_NA = TRUE
)

Arguments

metalyzer_se

A Metalyzer object

categorical

A column defining the categorical variable

groups

A vector of column names of aggregated_data to calculate the ANOVA group wise. If the column does not exists in aggregated_data it is automatically added from meta data. The default value is set to NULL, which uses the existing grouping of aggregated_data.

impute_perc_of_min

A numeric value below 1

impute_NA

Logical value whether to impute NA values

Value

A data frame containing the log2 fold change for each metabolite

Examples

metalyzer_se <- MetAlyzer_dataset(file_path = example_extraction_data())
metalyzer_se <- renameMetaData(
  metalyzer_se,
  Extraction_Method = "Sample Description"
)
# reduced to only 'Acylcarnitines' (first metabolic class) for simplicity
drop_vec = unique(metalyzer_se@elementMetadata$metabolic_classes)[2:24]
metalyzer_se <- filterMetabolites(
  metalyzer_se,
  drop_metabolites = drop_vec
)
metalyzer_se <- filterMetaData(
  metalyzer_se,
  Tissue == "Drosophila"
)
metalyzer_se <- calculate_anova(
  metalyzer_se,
  categorical = "Extraction_Method",
  groups = c("Metabolite"),
  impute_perc_of_min = 0.2,
  impute_NA = TRUE
)