noiseFixedEffects: Simulate noise fixed effects.

In HannahVMeyer/PhenotypeSimulator: Flexible Phenotype Simulation from Different Genetic and Noise Models

Description

noiseFixedEffects simulates a number of non-genetic covariate effects (confounders). Confounders can have effects across all traits (shared) or to a number of traits only (independent); in addition, only a certain proportion of traits can be affected by the confounders. Confounders can be simulated as categorical variables or following a binomial , uniform or normal distribution. Effect sizes for the noise effects can be simulated from a uniform or normal distribution. Multiple confounder sets drawn from different distributions/different parameters of the same distribution can be simulated by specifying NrFixedEffects and supplying the respective distribution parameters.

Usage

noiseFixedEffects(
  N,
  P,
  NrConfounders = 10,
  sampleID = "ID_",
  phenoID = "Trait_",
  id_samples = NULL,
  id_phenos = NULL,
  pTraitsAffected = 1,
  NrFixedEffects = 1,
  pIndependentConfounders = 0.4,
  pTraitIndependentConfounders = 0.2,
  keepSameIndependent = FALSE,
  distConfounders = "norm",
  mConfounders = 0,
  sdConfounders = 1,
  catConfounders = NULL,
  probConfounders = NULL,
  distBeta = "norm",
  mBeta = 0,
  sdBeta = 1,
  verbose = FALSE
)

Arguments

N	Number [integer] of samples to simulate.
P	Number [integer] of phenotypes to simulate.
NrConfounders	Vector of number(s) [integer] of confounders from a specified distribution to simulate.
sampleID	Prefix [string] for naming samples.
phenoID	Prefix [string] for naming traits.
id_samples	Vector of [NrSamples] sample IDs [string]; if not provided constructed by paste(sampleID, 1:N, sep="").
id_phenos	Vector of [NrTraits] phenotype IDs [string]; if not provided constructed by paste(phenoID, 1:P, sep="").
pTraitsAffected	Vector of proportion(s) [double] of traits affected by the confounders. For non-integer results of pTraitsAffected*P, the ceiling of the result is used.
NrFixedEffects	Number [integer] of different confounder effects to simulate; allows to simulate fixed effects from different distributions or with different parameters; if only one type of confounder distribution is wanted, set NrFixedEffects=1 and choose the number of confounders with eg NrConfounders=10.
pIndependentConfounders	Vector of proportion(s) [double] of confounders to have a trait-independent effect.
pTraitIndependentConfounders	Vector of proportion(s) [double] of traits influenced by independent confounder effects.
keepSameIndependent	[boolean] If set to TRUE, the independent genetic effects always influence the same subset of traits.
distConfounders	Vector of name(s) [string] of distribution to use to simulate confounders; one of "unif", "norm", "bin", "cat_norm", "cat_unif".
mConfounders	Vector of mean/midpoint(s) [double] of normal/uniform distribution for confounders.
sdConfounders	Vector of standard deviation(s)/distance from midpoint(s) [double] of normal/uniform distribution for confounders.
catConfounders	Vector of number(s) of confounder categories [integer]; required if distConfounders "cat_norm" or "cat_unif".
probConfounders	Vector of probability(s) [double] of binomial confounders (0/1); required if distConfounders "bin".
distBeta	Vector of name(s) [string] of distribution to use to simulate effect sizes of confounders; one of "unif" or "norm".
mBeta	Vector of mean/midpoint [double] of normal/uniform distribution for effect sizes of confounders.
sdBeta	Vector of standard deviation/distance from midpoint [double] of normal/uniform distribution for effect sizes of confounders.
verbose	[boolean] If TRUE, progress info is printed to standard out

Value

Named list of shared confounder effects (shared: [N x P] matrix), independent confoudner effects (independent: [N x P] matrix), the confounders labeled as shared or independent effect (cov: [NrConfounders x N] matrix) and the simulated effect sizes of the confounders (cov_effect: [P x NrConfounders] dataframe).

See Also

simulateDist

Examples

# fixed noise effect with default setting
noiseFE <- noiseFixedEffects(P=5, N=20)

# 1 categorical fixed noise effect with uniform distribution of the 
# categories
noiseFE_catUnif <- noiseFixedEffects(P=10, N=20, NrConfounders=1, 
distConfounders="cat_unif", catConfounders=3)

# 10 fixed noise effect with uniform distribution between 1 and 5 (3 +/- 2) 
# categories
noiseFE_uniformConfounders_normBetas <- noiseFixedEffects(P=10, N=20, 
NrConfounders=10, distConfounders="unif", mConfounders=3, sdConfounders=2, 
distBeta="norm",  sdBeta=2)

 # 4 fixed noise effect with binomial distribution with p=0.2 
noiseFE_binomialConfounders_uniformBetas <- noiseFixedEffects(P=10, N=20, 
NrConfounders=4, distConfounders="bin", probConfounders=0.2, distBeta="norm", 
sdBeta=2)

 # 2 fixed noise effect with 1 binomial confounders and 1 normally 
 # distributed confounder; the latter only affects 2 traits 
 noiseFE_binomialandNormalConfounders <- noiseFixedEffects(P=10, N=20, 
 NrFixedEffects=2, pTraitsAffected =c (1,0.2), NrConfounders=c(2,2), 
 distConfounders=c("bin", "norm"),  probConfounders=0.2)

HannahVMeyer/PhenotypeSimulator documentation built on July 19, 2021, 7:41 a.m.