iCARH.simulate: Simulates longitudinal data based on the iCARH model.

In iCARH: Integrative Conditional Autoregressive Horseshoe Model

Description

Simulates longitudinal data based on the iCARH model. Returns two types of datasets with relevant parameters (see below).

Usage

iCARH.simulate(
  Tp,
  N,
  J,
  P,
  K,
  path.names = NULL,
  path.probs = FALSE,
  pathway.perturb.ratio = 0.5,
  Ygroupeff = NULL,
  Zgroupeff = NULL,
  fe = 0,
  num.corr.y = 0,
  beta.val = NULL,
  sigma2 = 1,
  arz = 0.7,
  sdx = 0.01
)

Arguments

Tp	number of time points
N	number of samples (by default first N/2 controls and last N/2 cases)
J	number of metabolites
P	number of pathways (will probably change)
K	number of bacteria profiles (Y variables)
path.names	pathways to sample from as specified in KEGG. If not specified, path.probs will be considered.
path.probs	if TRUE, KEGG like density of pathways per metabolite is used to sample from. If scalar, path.probs is the expected ratio of metabolites in each pathway. Needs to be specified if path.names is not.
pathway.perturb.ratio	expected ratio of perturbed pathways
Ygroupeff	vector of 2xK variables (treatment effect on Y variables)
Zgroupeff	vector of 2 variables for treatment effect
fe	fixed effect
num.corr.y	number of correlated Y variables. The last num.corr.y will be highly correlated to the first num.corr.y variables
beta.val	beta values (regression coefficients) to sample from. Values will be randomly sampled if not specified.
sigma2	individual variance of metabolites
arz	autoregressive coefficient for treatment simulation
sdx	noise for autoregressive process, recommended value is 0.01

Value

list with the following objects :

XX	metabolomics data, X data
Y	additional omic data, Y data
Z	treatment
beta	effects of Y variables on X variables, column K+1 represents effect of treatment on X variables
pathways	pathway adjacency matrices
path.perturb	which pathways are perturbed?
phi	"spatial" dependence parameter, indicative of pathway perturbation
arx	autoregressive coefficients for X data
ary	autoregressive coefficients for Y data

Examples

 data.sim = iCARH.simulate(4, 8, 10, 2, 2, path.probs=0.3, Zgroupeff=c(0,4),
beta.val=c(1,-1,0.5, -0.5))

iCARH documentation built on Aug. 28, 2020, 1:10 a.m.