nlnet: Non-Linear Network reconstruction from expression matrix

In nlnet: Nonlinear Network, Clustering, and Variable Selection Based on DCOL

Description

Non-Linear Network reconstruction method

Usage

nlnet(input, min.fdr.cutoff=0.05,max.fdr.cutoff=0.2, conn.proportion=0.007, 
gene.fdr.plot=FALSE, min.module.size=0, gene.community.method="multilevel", 
use.normal.approx=FALSE, normalization="standardize", plot.method="communitygraph")

Arguments

input	the data matrix with no missing values.
min.fdr.cutoff	the minimun allowable value of the local false discovery cutoff in establishing links between genes.
max.fdr.cutoff	the maximun allowable value of the local false discovery cutoff in establishing links between genes.
conn.proportion	the target proportion of connections between all pairs of genes, if allowed by the fdr cutoff limits.
gene.fdr.plot	whether plot a figure with estimated densities, distribution functions, and (local) false discovery rates.
min.module.size	the min number of genes together as a module.
gene.community.method	the method for community detection.
use.normal.approx	whether to use the normal approximation for the null hypothesis.
normalization	the normalization method for the array.
plot.method	the method for graph and community ploting.

Details

gene.community.method: It provides three kinds of community detection method: "mutilevel", "label.propagation" and "leading.eigenvector".

use.normal.approx: If TRUE, normal approximation is used for every feature, AND all covariances are assumed to be zero. If FALSE, generates permutation based null distribution - mean vector and a variance-covariance matrix.

normalization: There are three choices: "standardize" means removing the mean of each row and make the standard deviation one; "normal_score" means normal score transformation; "none" means do nothing. In that case we still assume some normalization has been done by the user such that each row has approximately mean 0 and sd 1.

plot.method: It provides three kinds of ploting method: "none" means ploting no graph, "communitygraph" means ploting community with graph, "graph" means ploting graph, "membership" means ploting membership of the community

Value

it returns a graph and the community membership of the graph.

algorithm	The algorithm name for community detection
graph	An igraph object including edges : Numeric vector defining the edges, the first edge points from the first element to the second, the second edge from the third to the fourth, etc.
community	Numeric vector, one value for each vertex, the membership vector of the community structure.

Author(s)

Haodong Liu <[e-mail]>

References

https://www.ncbi.nlm.nih.gov/pubmed/27380516

See Also

data.gen

Examples

 
 ## generating the data matrix & hiden clusters as a sample
  input<-data.gen(n.genes=40, n.grps=4)
## now input includes data matrix and hiden clusters, so get the matrix as input.
input<-input$data 
##change the ploting method
 result<-nlnet(input,plot.method="graph")
  ## get the result and see it values
 graph<-result$graph ##a igraph object.
 comm<-result$community ##community of the graph
 
 ## use different community detection method
 #nlnet(input,gene.community.method="label.propagation")
 
 ## change the fdr pro to control connections of genes
 ## adjust the modularity size
 #nlnet(input,conn.proportion=0.005,min.module.size=10)
 

nlnet documentation built on Jan. 13, 2021, 10:35 a.m.