diffuse_mc: Compute the heatrank using permutations

In diffuStats: Diffusion scores on biological networks

Description

Function diffuse_mc has an implemented parallelisation of the Monte Carlo trials for diffusion in a network. The input scores are assumed to be sparse and are internally sparsified, so very dense scores migth take time with current implementation.

Usage

diffuse_mc(
    graph,
    scores,
    n.perm = 10000,
    sample.prob = NULL,
    seed = 1,
    oneminusHeatRank = TRUE,
    K = NULL,
    ...
)

Arguments

graph	igraph object
scores	Recursive list, can have either binary or quantitative scores
n.perm	Numeric, number of permutations
sample.prob	Numeric, probabilities (needn't be scaled) to permute the input. This is passed to sample's prob argument. If NULL, sampling is uniform. It has to be in a list format, with the same names as scores, and each element of the list must be the sampling probability of each background.
seed	Numeric, seed for random number generator
oneminusHeatRank	Logical, should 1 - heatrank be returned instead of heatrank?
K	Kernel matrix (if precomputed). If K is not supplied, the regularised Laplacian will be computed on the fly and used.
...	currently ignored arguments

Value

A list containing matrices of heatrank scores

Examples

# Using a list as input (needed)
data(graph_toy)
list_input <- list(myInput1 = graph_toy$input_mat)
diff_mc <- diffuse_mc(
    graph = graph_toy,
    scores = list_input)

diffuStats documentation built on Feb. 22, 2021, 10 a.m.