sim.phe: Simulation for phenotypes (SimPhe main process)

In SimPhe: Tools to Simulate Phenotype(s) with Epistatic Interaction

Description

Main process of simulation for phenotypes

Usage

sim.phe(sim.pars = NULL, fgeno = NULL, ftype = c("ind.head", "plink",
  "snp.head"), fwrite = TRUE, fphename = "simu.pheno",
  fusepar = "usedpars.txt", seed = NA, Dskim = 0, noise.var = 1,
  pattern = "[[:alpha:]]+", plink.path = system("which plink"),
  genetic.model = "epistasis", ...)

Arguments

sim.pars	a prepared list containing the parameters settings for simulation or a file of parameters settings. Please set your own parameters following the same structure as the object genepars or as the file simupars.txt (you could find example file system.file("extdata","simupars.txt",package="SimPhe")). To specify heritability, there are two ways: one is to set heritability in parameter file or in the prepared list object which will futher pass to sim.pars. Another way is to set noise.var by using function get.noise.var given specify heritability. List format: please follow the example of the object genepars. The meaning of each elment in the list is similar like the file format description below. File format: please follow the example of the simupars.txt file found in the inst/extdata/ directory of the package (run system.file("extdata", "simupars.txt", package="SimPhe") to to get the path to the file), blank lines are ignored. The file consists of three or four blocks for each phenotype (the number of blocks depends on user): mean, main and epistasis, sometimes heritability. Each block is started by a line of the form '[blockname]' followed by the parameter setting for the block, e.g. for first phenotype, [P1mean] mean β_0: coefficient parameter of "basic" genetic effects in Gij = β_0 + ∑(β_Gwt*wtij), t in (1, 2, ..., 8). [P1main] SNP SNP name additive coefficient of additive effect dominance coefficient of dominance effect [P1epistasis] SNPA first SNP SNPB second SNP additive_additive coefficient for additive-additive interaction additive_dominance coefficient for additive-dominance interaction dominance_additive coefficient for dominance-additive interaction dominance_dominance coefficient for dominance-dominance interaction [P1heritability] heritability expected heritability Similar meanings for "[P2mean]", "[P2main]", "[P2epistasis]", and so on.
fgeno	file to read genotype information from or pre-read data.frame with that information (matching the output format of read.geno).
ftype	genotype file format, it accepts three options: "plink": plink format (.bed, .bim, .fam or .map, .ped); "ind.head": columns are the indviduals and lines are SNPs; "snp.head": columns are SNPs and lines are indviduals. For "plink", fgeno needs to be given without suffix and plink.path may need to be assigned by the user because plink will be run from within SimPhe. More detail see plink.path. For the other options, fgeno should be the full name (with suffix and path if necessary) of the genotype file. Of course, this does not apply if fgeno is provided as a data frame.
fwrite	logical. Write out file (simulated data) or not. If TRUE (default), simulated phenotypes will be written, respectively.
fphename	filename of the phnotype(s). Default is "simu.pheno".
fusepar	filename of the setting for simulation (for recording). Default is "usedpars.txt".
seed	an integer used for set.seed(). Default is NA.
Dskim	the coefficient of linkage disequilibrium. Default is 0 (no LD).
noise.var	variance for random noise. Default is 1. Note that this is overridden by the heritability setting in the simulation parameter file. If heritability is given in parameter file then noise.var will not work.
pattern	ignore pattern for detecting the phenotype index from the parameter names. Default is "[[:alpha:]]+" which means letters.
plink.path	path of plink executable. Only needed when the ftype is "plink". Default is NULL. The function will detect the plink path with system("where plink") for Windows users and system("which plink") for Linux and MacOS users. But there is no garantee that the commands work on all devices. If the path cannot be determined or the executable cannot be called from read.geno, then users have to try other formats.
genetic.model	a string show the genetic model to use for simulation. Default is "epistasis".
...	not used.

Details

further discussion on pattern

Value

a data.frame with the simulated phenotype(s) where the column(s) refer to different phenotype(s) and rows to individuals.

Author(s)

Beibei Jiang [e-mail] and Benno Pütz [e-mail]

Examples

#### file path of example
# simulation parameters:
fpar.path <- system.file("extdata", "simupars.txt", package="SimPhe")

# genotype file: rows are individuals and columns are SNPs
fgeno.path <- system.file("extdata", "10SNP.txt", package="SimPhe")


#### instead of a parameter file, prepared list like genepars also works
genepars


#### simulate phenotype(s)
phe <- sim.phe(sim.pars = fpar.path, fgeno = fgeno.path, ftype = "snp.head", fwrite = FALSE)
# or
phe <- sim.phe(sim.pars = genepars, fgeno = fgeno.path, ftype = "snp.head", fwrite = FALSE)

# the simulated phenotype(s)
str(phe)
head(phe)

SimPhe documentation built on May 1, 2019, 9:10 p.m.