set.init: Set initial values for hyperparameters

In chenchunyu88/BATools: BATools: An R package that extends 'Bayesian Alphabet' for Whole Genome Prediction and Genome-wide association study

Description

Set initial values for hyperparameters

Usage

set.init(formula, data, geno, genoid, df = 5, scale = NULL, vare = NULL,
  g = NULL, beta = NULL, pi_snp = 0.05, post_prob = NULL, h2 = 0.5,
  c = 1000, model = c("GBLUP", "rrBLUP", "BayesA", "BayesB", "SSVS",
  "ssGBLUP", "ssBayesA", "ssBayesB", "ssSSVS", "anteBayesA", "anteBayesB"),
  varGenetic = NULL, centered = T, from = NULL)

Arguments

data	data.frame containing phenotypes
geno	genotype matrix
genoid	character of the column name of 'data' representing the id related the rowname of 'geno'
df	degrees of freedom parameter
scale	scale parameter, if is 'NULL' set automatically
vare	residual variance, if is 'NULL' set automatically
g	starting values for SNP effects, if is 'NULL' will start from
beta	starting values for fixed effects,
pi_snp	proportion of SNP that have non-zero effect for BayesB and SSVS, default is 0.05
h2	hertiability
c	ratio for the smaller variance for SSVS
model	the model used for analysis, can be c("rrBLUP","GBLUP", "BayesA","BayesB", "SSVS","ssGBLUP", "ssBayesA","ssBayesB", "ssSSVS","anteBayesA","anteBayesB")
centered	logical indicating if Z is centered
from	the source of scale, if it's 'rrBLUP', the scale will be adjusted for BayesA and SSVS
y	character phenotypes name

Value

a list of initial values

Examples

## Not run: 
 data(Pig)
 init=set_init(driploss~1,data=PigPheno,geno=geno,genoid = ~id,model="SSVS",centered=TRUE)

## End(Not run)

chenchunyu88/BATools documentation built on May 19, 2019, 8:21 a.m.