padma: Calculate individualized deviation scores from multi-omic...

In andreamrau/padma: Individualized Multi-Omic Pathway Deviation Scores Using Multiple Factor Analysis

Calculate individualized deviation scores from multi-omic data

Description

This is the primary user interface for the padma package. Generic S4 methods are implemented to calculate individualized pathway deviation scores on the basis of matched, multi-omic data. The supported classes for input are list and MultiAssayExperiment. The output of padma is an S4 object of class padmaResults.

Usage

padma(object, ...)

## S4 method for signature 'list'
padma(
  object,
  colData,
  gene_map = padma::mirtarbase,
  base_ids = NULL,
  supp_ids = NULL,
  pathway_name = "c2_cp_BIOCARTA_D4GDI_PATHWAY",
  impute = FALSE,
  variance_threshold = 1e-04,
  full_results = TRUE,
  verbose = TRUE,
  ...
)

## S4 method for signature 'MultiAssayExperiment'
padma(
  object,
  gene_map = padma::mirtarbase,
  base_ids = NULL,
  supp_ids = NULL,
  pathway_name = "c2_cp_BIOCARTA_D4GDI_PATHWAY",
  impute = FALSE,
  variance_threshold = 1e-04,
  full_results = TRUE,
  verbose = TRUE,
  ...
)

Arguments

object	Matched multi-omic data. May be provided as (1) a MultiAssayExperiment or (2) a named list, with each element corresponding to a matrix representing an omic, with biological entities in rows. Row names should include unique biological entity IDs (e.g., gene symbols, miRNA names); columns represent individuals. If more than one biological entity is used, a gene_map data.frame providing mappings between IDs and gene names should be provided if the default mirtarbase is not sufficient.
...	Optional additional arguments
colData	(optional) A DataFrame or data.frame of characteristics for all biological units, to be used in creating a MultiAssayExperiment from an object of class list
gene_map	(optional) Data frame mapping arbitrary biological entities (e.g. miRNAs) to genes. Contains two columns, where the first provides the IDs of the entity and the second provides the IDs of the corresponding target gene. By default, the miRNA-gene interactions of type 'Functional MTI' from miRTarBase are used (see the preloaded 'mirtarbase' data in the package).
base_ids	(optional) Sample names to be used as reference base data. By default, all samples are used.
supp_ids	(optional) Sample names to be used as supplementary individuals to be projected onto the analysis based on the individuals identified in base_ids. By default, takes the value NULL, but should not overlap with base_ids if provided by the user.
pathway_name	Character of either a KEGG pathway identifier or MSigDB pathway names (e.g., see the pathway names in the 'geneset' column of the preloaded msigdb data in the package), or a vector of gene symbols.
impute	If TRUE, impute missing values separately in base and supplementary data using MFA as implemented in the missMDA package; otherwise simple mean imputation is used (default).
variance_threshold	Minimal variance required across samples to retain a biological entity in the analysis
full_results	If TRUE (default), include full MFA results in function output; otherwise, provide concise output to save space.
verbose	If TRUE, provide verbose output.

Value

An S4 object of class padmaResults, where individualized pathway deviation scores are stored as the assay data, and the corresponding {pathway name, full MFA results, number of genes, and names of imputed or filtered genes} are stored as slots that can be retrieved using the appropriate accessor functions.

Author(s)

Andrea Rau

Examples

LUAD_subset <- padma::LUAD_subset
## Create MultiAssayExperiment object with LUAD data
omics_data <- 
  list(rnaseq = as.matrix(LUAD_subset$rnaseq),
       methyl = as.matrix(LUAD_subset$methyl),
       mirna = as.matrix(LUAD_subset$mirna),
       cna = as.matrix(LUAD_subset$cna))
pheno_data <- 
  data.frame(LUAD_subset$clinical, 
             row.names = LUAD_subset$clinical$bcr_patient_barcode)
mae <-
  suppressMessages(
    MultiAssayExperiment::MultiAssayExperiment(
      experiments = omics_data, colData = pheno_data))

## Run padma
run_padma <- 
  padma(mae, gene_map = padma::mirtarbase,
        pathway_name = "c2_cp_BIOCARTA_D4GDI_PATHWAY", verbose = FALSE)

summary(run_padma)

## padma plots
## Not run: 
factorMap(run_padma, dim_x = 1, dim_y = 2)
factorMap(run_padma, dim_x = 1, dim_y = 2,
           partial_id = "TCGA-78-7536")
omicsContrib(run_padma, max_dim = 10)

## End(Not run)

andreamrau/padma documentation built on June 15, 2024, 8:39 a.m.