kobas.gsea {KOBAS} R Documentation

Run the KOBAS style GSEA analysis

Description

<br /> Gene Set Enrichment Analysis with the KOBAS ranking and permutation scheme: the<br /> genes are ranked by the signal-to-noise ratio or the t statistics between the<br /> two phenotype groups, then a weighted Kolmogorov-Smirnov like running sum is<br /> evaluated for each gene set. The significance of the enrichment score is<br /> estimated by permuting the phenotype class labels.

Usage

kobas.gsea(expr, cls, gmt,
    method = "snr",
    weight = 1,
    permutations = 1000,
    min.size = 15,
    max.size = 500,
    seed = 20261026,
    keep.running.sum = FALSE,
    verbose = TRUE);

Arguments

expr

the expression matrix, rows are genes and columns are samples. The value could be a matrix object, a dataframe object (rows are genes and columns are samples), or a file path of the gct / KOBAS plain text expression matrix.

cls

the phenotype class labels: could be a phenotype object, a CLS file path, an integer vector (0 for the first phenotype group and 1 for the second group), or a character vector with exactly two distinct class labels.

gmt

the gene set collection: a Background model or a GMT file path.

method

the gene ranking method: snr (signal to noise ratio, default) or ttest. [as string]

weight

the weight of the enrichment score, 0 for the unweighted Kolmogorov-Smirnov statistic, 1 (default) for the classic GSEA weighting and 2 for the over-weighted score. Any other positive value p will use |rank|^p. [as double]

permutations

the number of the phenotype label permutations, default 1000. Set to 0 for skipping the significance estimation (then the pvalue and FDR columns of the result table will be NA and NES equals to ES). [as integer]

min.size

the minimum number of the mapped members of a gene set, default 15. [as integer]

max.size

the maximum number of the mapped members of a gene set, default 500. [as integer]

seed

the random seed of the permutation, default 20261026. [as integer]

keep.running.sum

keeps the running sum of each gene set in the result objects? Default FALSE because the running sum occupies a lot of memory for large gene set collections. [as boolean]

verbose

prints the progress message? Default TRUE. [as boolean]

env

the R# runtime environment. [as Environment]

Details

A positive ES means that the gene set is enriched in the first phenotype group (the samples labeled with 0) because the KOBAS ranking score is mean(group0) - mean(group1).

Authors

gseakit

Value

a dataframe with the gene set id as the row names and the columns: name, description, size, ES, NES, pvalue, FDR, es_index, leading_edge (the hit genes before/after the enrichment score position, joined by ;) and hit_genes.

clr value class

Examples

[Package KOBAS version 1.0.0.0 Index]