gnet: Run GNET2

In jianlin-cheng/GNET2: Constructing gene regulatory networks from expression data through functional module inference

Description

Build regulation modules by iteratively perform regulator assigning and Gene assigning, until the assignment of genes did not change, or max number of iterations reached.

Usage

gnet(
  input,
  reg_names,
  init_method = "boosting",
  init_group_num = 4,
  max_depth = 3,
  cor_cutoff = 0.9,
  min_divide_size = 3,
  min_group_size = 2,
  max_iter = 5,
  heuristic = TRUE,
  max_group = 0,
  force_split = 0.5,
  nthread = 4
)

Arguments

input	A SummarizedExperiment object, or a p by n matrix of expression data of p genes and n samples, for example log2 RPKM from RNA-Seq.
reg_names	A list of potential upstream regulators names, for example a list of known transcription factors.
init_method	Cluster initialization, can be "boosting" or "kmeans", default is using "boosting".
init_group_num	Initial number of function clusters used by the algorithm.
max_depth	max_depth Maximum depth of the tree.
cor_cutoff	Cutoff for within group Pearson correlation coefficient, if all data belong to a node have average correlation greater or equal to this, the node would not split anymore.
min_divide_size	Minimum number of data belong to a node allowed for further split of the node.
min_group_size	Minimum number of genes allowed in a group.
max_iter	Maxumum number of iterations allowed if not converged.
heuristic	If the splites of the regression tree is determined by k-means heuristicly.
max_group	Max number of group allowed for the first clustering step, default equals init_group_num and is set to 0.
force_split	Force split the largest gene group into smaller groups by kmeans. Default is 0.5(Split if it contains more than half target genes)
nthread	Number of threads to run GBDT based clustering

Value

A list of expression data of genes, expression data of regulators, within group score, table of tree structure and final assigned group of each gene.

Examples

set.seed(1)
init_group_num = 8
init_method = 'boosting'
exp_data <- matrix(rnorm(50*10),50,10)
reg_names <- paste0('TF',1:5)
rownames(exp_data) <- c(reg_names,paste0('gene',1:(nrow(exp_data)-length(reg_names))))
colnames(exp_data) <- paste0('condition_',1:ncol(exp_data))
se <- SummarizedExperiment::SummarizedExperiment(assays=list(counts=exp_data))
gnet_result <- gnet(se,reg_names,init_method,init_group_num)

jianlin-cheng/GNET2 documentation built on Oct. 22, 2020, 11:24 p.m.