rwr_OVprop: Network based propagation with random walk for the multi-gene...

In LucaGiudice/Esearch3D: Leveraging gene expression and chromatin architecture to predict intergenic enhancers

Description

This function performs propagation of each individual cell expression profile associated to a column of an expression matrix. The cell expression profile is characterized by the expression of genes at the rows. The values of the genes indicate their expression in a specific column and cell profile The propagation considers each cell profile at a time. It maps the expression values to the corresponding genes represented as nodes in the network. It diffuses the gene expression values. Each node in the CIN network composed by genes, genic fragments and intergenic fragments gets a propagation score which measures its activity. The imputed activity score (IAS) obtained by genes and fragments replaces their original value in the cell expression profile. The result is a new matrix of cell profiles. Each cell profile is now described with genes and also fragments who own a IAS.

Usage

rwr_OVprop(
  g,
  input_m,
  norm = "column",
  no_cores = 2,
  r = 0.8,
  stop_step = 200,
  stop_delta = 1e-06,
  keep_zero = F
)

Arguments

g	igraph, edge list or adjacency matrix of the chromatin interaction network (CIN)
input_m	numeric matrix of cell profiles, columns are cells, rows are genes or fragments, value is expression of IAS
norm	Default column, options: row or column or laplacian, it sets the graph normalization performed on the network
no_cores	Default 2, An integer value greater than 0 indicating the number of cores to use for computing in parallel the task
r	Default 0.8, A double value lower than 1 indicating the percentage of information that a gene keeps (0.8 is 80 percentage)
stop_step	Default 200, An integer value greater than 0 indicating the number of iterations of the propagation
stop_delta	Default 1e-06, A double value lower than 0 indicating the threshold under which all imputed propagation values are set 0
keep_zero	Default FALSE, Boolean: TRUE if to keep profiles without information

Value

matrix of propagated profiles with imputed activity scores for the genes and fragments in the network and that replace the original expression profiles

LucaGiudice/Esearch3D documentation built on Dec. 17, 2021, 1:12 a.m.