run_entire_gwas_pipeline: MLMM, model selection and effects estimation

In mlmm.gwas: Pipeline for GWAS Using MLMM

Description

Internaly run functions of the mlmm.gwas package:

• mlmm_allmodels (GWAS)

• frommlmm_toebic

• eBIC_allmodels (model selection)

• threshold_allmodels

• fromeBICtoEstimation

• Estimation_allmodels (effects estimation)

Usage

run_entire_gwas_pipeline(Y, XX, KK, nbchunks = 2, maxsteps = 20,
  cofs = NULL, female = NULL, male = NULL,threshold=NULL,lambda=NULL)

Arguments

Y	A numeric named vector where the names are individuals' names and the values their phenotype. The names of Y will be matched to the row names of X.
XX	A list of length one, two or three matrices depending on the model. Matrices are n by m matrix, where n=number of individuals, m=number of SNPs, with rownames(X)=individual names, and colnames(X)=SNP names. - additive: a single matrix - additive+dominance: two matrices - female+male: two matrices with the female one first - female+male+interaction: three matrices with the female first, the male then the interaction
KK	a list of one, two or three matrices depending on the models - additive: a n by n matrix, where n=number of individuals, with rownames()=colnames()=individual names - additive+dominance: two n by n matrices, where n=number of individuals, with rownames()=colnames()=individual names - female+male: a n.female by n.female matrix, with rownames()=colnames()=female names and a n.male by n.male matrix, with rownames()=colnames()=male names - female+male+interaction: the same two matrices as the model female+male and a n by n matrix, where n=number of individuals, with rownames()=colnames()=individual names
nbchunks	An integer defining the number of chunks of matrices to run the analysis, allows to decrease the memory usage. minimum=2, increase it if you do not have enough memory
maxsteps	An integer >= 3. Maximum number of steps desired in the forward approach. The forward approach breaks automatically once the pseudo-heritability is close to 0, however to avoid doing too many steps in case the pseudo-heritability does not reach a value close to 0, this parameter is also used.
cofs	A n by q matrix, where n=number of individuals, q=number of fixed effect, with rownames()=individual names and with column names, forbidden head of column names for this matrix "eff1_" and usage of special characters as "*","/","&"
female	A factor of levels female names and length n, only for the last two models
male	A factor of levels male names and length n, only for the last two models
threshold	a value to declare the significant p value. The default value is Bonferroni 0.05
lambda	penalty used in the computation of the eBIC; if NULL, the default will be 1 - 1/(2k) with L=n^k where L=total number of SNPs (see function "lambda.calc")

Value

A named list with 2 or 3 elements:

• pval: the return value of mlmm_allmodels

• eBic: the return value of eBIC_allmodels

• threshold: the return value of threshold_allmodels

• effects: the return value of Estimation_allmodels, only if there is at least one marker in the model selected by lowest eBIC.

Examples

data("mlmm.gwas.AD")
results <- run_entire_gwas_pipeline(floweringDateAD, list(Xa), list(K.add))

mlmm.gwas documentation built on Aug. 5, 2019, 5:12 p.m.