samplephi: sample random initial networks

In ddepn: Dynamic Deterministic Effects Propagation Networks: Infer signalling networks for timecourse RPPA data.

Description

Internal function.

Usage

samplephi(phi,stimuli, antibodies, tps, reps, dat, searchstatespace=FALSE,
		hmmiterations=5, phiasis=FALSE, lambda=NULL, B=NULL, Z=NULL, fanin=4,
		gam=NULL, it=NULL, K=NULL, priortype="none", scale_lik=FALSE,
		 allow.stim.off=TRUE, implementation="C")
initialphi(dat, phi, stimuli, Lmax, thetax, gammax, gammaposs,
		tps, reps, antibodies, n=100, multicores=FALSE, lambda=NULL, 
		B=NULL, Z=NULL, gam=NULL, it=NULL, K=NULL, priortype="none", 
		scale_lik=FALSE, allow.stim.off=TRUE, implementation="C")

Arguments

dat	The data matrix
phi	The initial network. Can be NULL.
stimuli	The stimuli list.
hmmiterations	Integer specifying the maximum number of iterations in the statespacesearch
multicores	TRUE for using multiple cores.
Lmax	Likelihood score
thetax	Parameter matrix
gammax	Statespacematrix
gammaposs	All possible state vectors
tps	The timepoints
reps	Number of replicates
antibodies	Character vector of protein names in the network.
searchstatespace	Do statespacesearch in initial sampling?
phiasis	TRUE for taking a given phi matrix as sample
lambda	The Prior influence strength in the laplace prior.
B	The Prior information matrix. Corresponds to prior edge probabilities in the final network.
Z	The normalisation factor for the prior distribution.
fanin	Integer: maximal indegree for nodes.
gam	Prior influence strength in scalefree prior.
it	Number of iterations to generate background distribution in scalefree prior.
k	Scaling factor in scalefree prior
priortype	Character. Type of prior to be used.
scale_lik	Boolean.Scale likelihood according to number of data points from which the overall likelihood is calculated.
allow.stim.off	Boolean. Allow the stimuli to become inactive at some point. See also ddepn.
implementation	String. One of "C","R","R_globalest","C_globalest". Different implementations of the HMM in perform.hmmsearch. If "R", the original pure R-implementation is used, if "C", a ported C-implementation is used. If "R_globalest", an experimental version of the parameter estimation is used in the HMM, "C_globalest" is the C-port of this version. See details for a description.

Details

Used for initializing random networks. Called internally.

Value

TODO

Note

TODO

Author(s)

Christian Bender

References

TODO

See Also

TODO

Examples

##TODO

ddepn documentation built on May 2, 2019, 4:42 p.m.