moduleNetwork {nem} | R Documentation |
Function moduleNetwork
estimates the hierarchy using a divide and conquer approach. In each step only a subset of nodes (called module)
is involved and no exhaustive enumeration of model space is needed as in function score
.
moduleNetwork(D,type="mLL",Pe=NULL,Pm=NULL,lambda=0,delta=1,para=NULL,hyperpara=NULL,verbose=TRUE) ## S3 method for class 'ModuleNetwork': print(x,...)
D |
data matrix. Columns correspond to the nodes in the silencing scheme. Rows are phenotypes. |
type |
see nem |
Pe |
prior position of effect reporters. Default: uniform over nodes in hierarchy |
Pm |
prior on model graph (n x n matrix) with entries 0 <= priorPhi[i,j] <= 1 describing the probability of an edge between gene i and gene j. |
lambda |
regularization parameter to incorporate prior assumptions. |
delta |
regularization parameter for automated E-gene subset selection (CONTmLLRatio only) |
para |
vector with parameters a and b for "mLL", if count matrices are used |
hyperpara |
vector with hyperparameters a0, b0, a1, b1 for "FULLmLL" |
verbose |
do you want to see progress statements printed or not? Default: TRUE |
x |
nem object |
... |
other arguments to pass |
moduleNetwork
is an alternative to exhaustive search
by the function score
and more accurate than pairwise.posterior
and triples.posterior
.
It uses clustering to sucessively split the network into smaller modules, which can then be estimated completely. Connections between modules are estimated by performing a constraint greedy hillclimbing.
nem object
Holger Froehlich
data("BoutrosRNAi2002") res <- nem(BoutrosRNAiDiscrete[,9:16],para=c(.13,.05),inference="ModuleNetwork") # plot graph plot(res,what="graph") # plot posterior over effect positions plot(res,what="pos") # estimate of effect positions res$mappos