mdp: Molecular Degree of Perturbation

In melissalever/mdp: Molecular Degree of Perturbation calculates scores for transcriptome data samples based on their perturbation from controls

Description

Based on the Molecular Distance to Health, this function calculates scores to each sample based on their perturbation from healthy

Usage

mdp(data, pdata, control_lab, directory = "", pathways, print = TRUE,
  measure = c("mean", "median"), std = 2, fraction_genes = 0.25,
  save_tables = TRUE, file_name = "")

Arguments

data	data frame of gene expression data with the gene symbols in the row names
pdata	data frame of phenodata with a column headed Class and the other headed Sample.
control_lab	character vector specifying the control class
directory	(optional) character string of output directory
pathways	(optional) list whose names are pathways and elements are genes in the pathway. see details section for more information
print	set as default to TRUE for pdfs of the sample scores to be saved
measure	'medan' as default, can change to 'median'. mean will select for z-score and median will select for modified z-score. (see details)
std	numeric set as default to 2, this governs the thresholding of expression data. z-scored expression values with absolute value less than 'std' will be set to 0.
fraction_genes	numeric fraction of genes that will contribute to the top perturbed genes. Set as default to 0.25
save_tables	Set as default to TRUE. Tables of zscore and gene and sample scores will be saved.
file_name	(optional) character string that will be added to the saved file names

Value

A list: zscore, gene_scores, gene_freq, sample_scores, perturbed_genes

• Z-score - z-score is calculated using the control samples to compute the average and the standard deviation. The absolute value of this matrix is taken and values less than the std are set to zero. This z-score data frame is used to compute the gene and sample scores.

• Gene scores - mean z-score value for each gene in each class

• Gene frequency - frequency with which a gene has a non zero z-score value in each class

• Sample scores - list containing sample scores for different genesets. Sample scores are the sum of the z-scored gene values for each sample, averaged for the number of genes in that geneset.

• Perturbed genes - vector of the top fraction of genes that have higher gene scores in the test classes compared to the control.

• Pathways - if genesets are provided, they are ranked according to the signal-to-noise #' ratio of test sample scores versus control sample scores calculated using that geneset.

Loading pathways

a list of pathways can be loaded from a .gmt file using the fgsea function using fgsea::gmtPathways('gmt.file.location')

Selecting mean or median

if median is selected, the z-score will be calculated using the median, and the standard deviation will be estimated using the median absolute deviation, utilising the mad function.

Examples

# basic run
mdp(example_data,example_pheno,'baseline')
# run with pathways
pathway_file <- system.file('extdata', 'ReactomePathways.gmt', 
package = 'mdp')
mypathway <- fgsea::gmtPathways(pathway_file) # load a gmt file
mdp(data=example_data,pdata=example_pheno,control_lab='baseline',
pathways=mypathway)

melissalever/mdp documentation built on Aug. 8, 2021, 10:45 p.m.