BayesCpi: 'BayesCpi model for GS and GWAS'

In jyanglab/g3tools: For Genomics, quantGen, and popGen studies, especially for crop species.

Description

R codes based on Rohan Fernando and Dorian Garrick, Bayesian Methods Applied to GWAS. All rights and credits go to Rohan Fernando and Dorian Garrick. The code is not efficient for large scale genomic data.

Usage

BayesCpi(genofile = "data/bayes_geno.txt",
  train_pheno = "data/bayes_train_pheno.txt", test_pheno = NULL,
  mergeby = "uid", trait = "TA", seed = 12347, chainLength = 1000,
  probFixed = 0.999, estimatePi = "yes", dfEffectVar = 4, nuRes = 4,
  varGenotypic = 1, varResidual = 1, windowSize = 10,
  outputFrequency = 100)

read_geno_pheno(genofile = "data/bayes_geno.txt",
  train_pheno = "data/bayes_train_pheno.txt", test_pheno = NULL, mergeby)

Arguments

genofile	A data.frame contains snpid information, must contain col: mergeby. [data.frame, ["uid", ...]].
train_pheno	A training data.frame contains columns of phenotypic traits and fixed effects. Can be used for GWAS. [data.frame, ["uid", "pheno", ...]].
test_pheno	A testing data.frame contains columns of phenotypic traits and fixed effects. For GS. [data.frame, ["uid", "pheno", ...], =NULL].
mergeby	The column id to merge the pheno and geno data. [chr, ="uid"]
seed	Seed for the random number generator. [integer, =12347].
chainLength	Number of iterations for the MCMC chain. [interger, =11000].
probFixed	Parameter "pi" the probability SNP effect is zero [num, =.99].
estimatePi	Wether to estimate pi or not. [chr, ="yes"].
dfEffectVar	Hyper parameter (degrees of freedom) for locus effect variance. [integer, =4].
nuRes	Hyper parameter (degrees of freedom) for residual variance. [integer, =4].
varGenotypic	Used to derive hyper parameter (scale) for locus effect variance. [num, =1].
varResidual	Used to derive hyper parameter (scale) for residual variance. [num, =1].
windowSize	Number of consecutive markers in a genomic window. [integer, =10].
outputFrequency	Frequency for reporting performance and for computing genetic variances. [integer, =100]

Value

return GWAS results.

jyanglab/g3tools documentation built on May 20, 2019, 6:27 a.m.