Description Usage Arguments Value Author(s) References See Also Examples
This function generates Reversible Jump MCMC (RJ-MCMC) sampling for approximating the posterior distribution of a time varying regulatory network, under the Auto Regressive TIme VArying (ARTIVA) model (for a detailed description of the algorithm, see Lebre et al. BMC Systems Biology, 2010).
Starting from time-course gene expression measurements for a gene of interest (referred to as "target gene") and a set of genes (referred to as "parent genes") which may explain the expression of the target gene, the ARTIVA procedure identifies temporal segments for which a set of interactions occur between the "parent genes" and the "target gene". The time points that delimit the different temporal segments are referred to as changepoints (CP).
If the measurements time delay is short enough so that the expression of
the target gene depends more likely on the expression of the parent
genes at some previous time points, then the time delay for the
interactions can be chosen with argument dyn
. In that case the set of
parent genes may contain the target gene (auto-regulation). Otherwise,
contemporaneous measurements of the parent genes are used to explain the
expression of the target gene and argument dyn
is set to 0. In the
latter case, the target gene must be kept out of the set of possible
parent genes.
The ARTIVA algorithm uses a combination of efficient and robust methods: (1) dynamical Bayesian networks (DBN) to model directed regulatory interactions between the parent genes and the analyzed target gene (2) RJ-MCMC sampling for inferring - simultaneously - (a) the time position (changepoints) at which the regulatory interactions between the parent genes and the target gene changes and (b) the regulatory network topologies (interactions between parent and target genes) associated with the temporal segments delimited by the changepoints.
If available, repeated measurements can be used, the design of
experiments must be specified with argument dataDescription
.
1 2 3 4 5 6 7 8 | ARTIVAsubnet(targetData, parentData, targetName="Target",parentNames=NULL,
dataDescription=NULL, saveEstimations=TRUE, saveIterations=FALSE,
savePictures = TRUE, outputPath=NULL, dyn=1, segMinLength=2, maxCP=NULL,
maxPred=NULL, nbCPinit=NULL, CPinit=NULL, niter=50000, burn_in=NULL,
PSRFactor=FALSE, PSRF_thres=1.05, segmentAnalysis=TRUE, edgesThreshold=0.5,
layout = "fruchterman.reingold", cCP= 0.5, cEdges=0.5, alphaCP=1,
betaCP=0.5, alphaEdges=1, betaEdges=0.5, v0=1, gamma0=0.1, alphad2=2,
betad2=0.2, silent=FALSE)
|
targetData |
A vector with the temporal gene expression measurements for the target gene (i.e. the gene whose regulation factors are looked for). |
parentData |
A matrix (or a vector if only 1 parent gene) with the temporal gene expression measurements for the proposed parent genes (i.e. potential regulation factors). Parent genes are shown in row and expression values in column. For computational reasons, we advise not to test simultaneously more than 10 parent genes. |
targetName |
Name of the target gene (optional, default: targetName="Target"). |
parentNames |
A vector with the names for parent gene(s) (optional, default:
|
dataDescription |
(Required only when the gene expression measurements contain repeated
values for the same time points). A vector indicating the ordering of
the time measurements in the data. For example
dataDescription=rep(1:n, each=2), if there are two measurements for
each time point AND the repetitions for each time point are next to
each other. Note that temporal gene expression measurements have to be
organized identically in arguments |
saveEstimations |
Boolean, if |
saveIterations |
Boolean, if |
savePictures |
Boolean, if |
outputPath |
File path to a folder in which the output results have to be saved,
either a complete path or the name of a folder to be created in the
current directory (optional, default: |
dyn |
Time delay to be considered in the auto-regressive process, between the
temporal expression measurements of the analyzed target gene and the
ones of the parent genes (optional, default: |
segMinLength |
Minimal length (number of time points) to define a temporal
segment. Must be - strictly - greater than 1 if there is no
repeated measurements for each time point in arguments
|
maxCP |
Maximal number of CPs to be considered by the ARTIVA inference
procedure. Note that for a temporal course with |
maxPred |
Maximal number of simultaneous incoming edges for each segment of the
regulatory network estimated for the target gene (default:
|
nbCPinit |
Number of CPs to be considered at the algorithm initialization
(optional, default: |
CPinit |
A vector with the initial positions for potential CPs.
(optional, default: |
niter |
Number of iterations to be performed in the RJ-MCMC sampling
(optional, default: |
burn_in |
Number of initial iterations that are discarded for the estimation of
the model distribution (posterior distribution). The
|
PSRFactor |
Boolean, if |
PSRF_thres |
(Only when |
segmentAnalysis |
Boolean, if |
edgesThreshold |
Probability threshold for the selection of the edges of the time-varying
regulatory network when |
layout |
Name of the function determining the placement of the vertices for
drawing a graph, possible values among others:
|
cCP |
Maximal probability to propose the birth (resp. death) of a changepoint
(CP) during the RJ-MCMC iterations (optional, default: |
cEdges |
Maximal probability - when a move update of the network topology is
chosen - to propose ecah edge move (birth or death of an edge) within a
temporal segment (optional, default: |
alphaCP |
Hyperparameter for sampling the number |
betaCP |
Hyperparameter for sampling the number |
alphaEdges |
Hyperparameter for sampling the number |
betaEdges |
Hyperparameter for sampling the number |
v0 |
Hyperparameter for sampling the noise variance (denoted by |
gamma0 |
Hyperparameter for sampling the noise variance (denoted by |
alphad2 |
Hyperparameter for sampling a parameter that represents the expected
signal-to-noise ratio (denoted by |
betad2 |
Hyperparameter for sampling a parameter that represents the expected
signal-to-noise ratio (denoted by |
silent |
Boolean, if |
Samples |
Results obtained at each iteration of the RJ-MCMC procedure. 1) 2) 3) 4) |
Counters |
Results obtained at each iteration of the RJ-MCMC procedure. 1) 2) 3) 4) 5) 6) |
CPpostDist |
A list of 2 tables : 1) 2) |
nbSegs |
An integer equal to the number of temporal segments with the largest value observed in the posterior distribution (see previously |
SegmentPostDist |
(only when parameter
1) 2) 3) |
network |
A table containing the information to plot (see function
|
GLOBvar |
A list of parameters used in the |
HYPERvar |
A list of hyperparameters used in the |
OUTvar |
A list of output parameters used in the |
targetData |
|
parentData |
|
S. Lebre and G. Lelandais
S. Lebre, J. Becq, F. Devaux, M. P. H. Stumpf, G. Lelandais (2010) Statistical inference of the time-varying structure of gene-regulation networks BMC Systems Biology, 4:130.
Gelman, A. and D. Rubin (1992) Inference from iterative simulation using multiple sequences. Statistical science 7 (4), 457-472.
ARTIVAnet
,ARTIVAsubnetAnalysis
,
choosePriors
, CP.postDist
, plotCP.postDist
,
segmentModel.postDist
, traceNetworks
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 | # Load the ARTIVA R package
library(ARTIVA)
# Load the dataset with simulated gene expression profiles
data(simulatedProfiles)
# Name of the target gene to be analyzed with ARTIVA
targetGene = 1
# Names of the parent genes (typically transcription factors)
parentGenes = c("TF1", "TF2", "TF3", "TF4", "TF5")
###
# ARTIVA analysis searching for potential interactions between the target
# genes and a predefined list of parent genes.
###
# Note that the number of iterations in the RJ-MCMC sampling is reduced
# to in this example to 'niter=20000' (in order obtain a quick overview
# of the ARTIVAnet fonction, but it should be increased (e.g. up to
# 50000) for a better parameter estimation.
# Without saving the output pictures "savePictures=FALSE"
## Not run:
ARTIVAtest = ARTIVAsubnet(targetData = simulatedProfiles[targetGene,],
parentData = simulatedProfiles[parentGenes,],
targetName = targetGene,
parentNames = parentGenes,
segMinLength = 2,
edgesThreshold = 0.5,
niter = 20000,
savePictures=FALSE)
# By default, the output results (pictures and estimation values) are
# saved in a folder named "ARTIVAsubnet" created in the current directory
ARTIVAtest = ARTIVAsubnet(targetData = simulatedProfiles[targetGene,],
parentData = simulatedProfiles[parentGenes,],
targetName = targetGene,
parentNames = parentGenes,
segMinLength = 2,
edgesThreshold = 0.5,
niter = 20000)
# In order to save the results in a specific folder, for example a
# folder entitled "ARTIVA-test" in the current directory:
ARTIVAtest2 = ARTIVAsubnet(targetData = simulatedProfiles[targetGene,],
parentData = simulatedProfiles[parentGenes,],
targetName = targetGene,
parentNames = parentGenes,
segMinLength = 2,
edgesThreshold = 0.5,
niter = 20000,
outputPath = "ARTIVA-test")
## End(Not run)
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.