crosstalkr (version 1.0.4)

sparseRWR: Perform random walk with repeats on a sparse matrix

Description

This function borrows heavily from the RWR function in the RANKS package (cite here)

Usage

sparseRWR(seed_proteins, w, gamma = 0.6, eps = 1e-10, tmax = 1000, norm = TRUE)

Value

numeric vector, affinity scores for all nodes in graph relative to provided seeds

Arguments

seed_proteins

user defined seed proteins

w

The adjacency matrix of a given graph in sparse format - dgCMatrix

gamma

restart probability

eps

maximum allowed difference between the computed probabilities at the steady state

tmax

the maximum number of iterations for the RWR

norm

if True, w is normalized by dividing each value by the column sum.

Examples

# 1) Run Random walk with restarts on a simple matrix
v1 = (c(1,1,1,0))
v2 = c(0,0,0,1)
v3 = c(1,1,1,0)
v4 = c(0,0,0,1)
w = matrix(data = c(v1,v2,v3,v4), ncol = 4, nrow = 4)
sparseRWR(seed_proteins = c(1,3), w = w, norm = TRUE)

# 2) Works just as well on a sparse matrix
v1 = (c(1,1,1,0))
v2 = c(0,0,0,1)
v3 = c(1,1,1,0)
v4 = c(0,0,0,1)
w = matrix(data = c(v1,v2,v3,v4), ncol = 4, nrow = 4)
w = Matrix::Matrix(w, sparse = TRUE)
sparseRWR(seed_proteins = c(1,4), w = w, norm = TRUE)
# \donttest{
#3) Sample workflow for use with human protein-protein interaction network
#g <- prep_biogrid()
#w <- igraph::as_adjacency_matrix(g)
#sparseRWR(seed_proteins = c("EGFR", "KRAS"), w = w, norm = TRUE)
# }