farmcpu: Perform GWAS using the FarmCPU model.

Description Usage Arguments Value Author(s)

View source: R/farmcpu.R

Description

Perform GWAS using the FarmCPU model.

Usage

1
2
3
4
5
6
farmcpu(Y, GD, GM, CV = NULL, GP = NULL, method.sub = "reward",
  method.sub.final = "reward", method.bin = "static",
  bin.size = c(5e+05, 5e+06, 5e+07), bin.selection = seq(10, 100, 10),
  memo = NULL, Prior = NULL, ncores.glm = 1, maxLoop = 10,
  converge = 1, iteration.output = FALSE, p.threshold = NA,
  ncores.reml = 1, threshold = 0.7)

Arguments

Y

A dataframe with two columns: a character column of sample names and a numeric column of phenotypic values. This dataframe may contain missing values.

GD

A big.matrix object.

GM

A dataframe with three columns: a character column of sample names, a numeric column of chromosomes, and a numeric column of base-pair positions.

CV

An optional numeric matrix of covariates.

GP

A numeric vector of p-values to select pseudo-QTNs before single-marker regression in the first iteration.

method.sub

Character string indicating the method used to substitute p-values for pseudo-QTNs. One of c('reward', 'mean', 'median', 'penalty', 'onsite').

method.sub.final

Character string indicating the method used to substitute p-values for pseudo-QTNs on the final iteration.

method.bin

Character string indicating the method used to select pseudo-QTNs. One of c('static', 'optimum').

bin.size

A numeric vector specifying bin sizes in base-pairs for pseudo-QTN selection.

bin.selection

Numeric vector of numbers of pseudo-QTNs to select.

memo

Character string added to file names.

Prior

A dataframe with the same format as GM with a fourth numeric column containing prior marker probabilities.

ncores.glm

Integer scalar number of cores to use for single-marker regression.

maxLoop

Integer scalar maximum number of iterations to perform.

converge

Numeric scalar such that 0 < converge < 1. Controls the percentage of overlapping pseudo-QTNs between two iterations required to terminate the GWAS.

iteration.output

Logical scalar. Whether to include the output of each iteration or not.

p.threshold

Numeric scalar. P-value threshold for inclusion of pseudo-QTNs in the model on the first iteration. Defaults to a 0.01 Bonferroni correction.

ncores.reml

Integer scalar number of cores to use for pseudo-QTN selection.

threshold

Numeric scalar maximum correlation allowed between pseudo-QTNs.

Value

A list. If iteration.output = FALSE, the list has one element. Otherwise, the list additionally has one element for each iteration. Each element has one required element and one optional element. The required element is a dataframe called GWAS with seven columns: the marker ID, the chromosome, the base-pair position, the marker p-value, the effect estimate, the standard error, and the t-statistic. If there are user-specified covariates, the results list contains a matrix called CV that contains the effect estimates for the covariates.

Author(s)

Xiaolei Liu

Zhiwu Zhang

Aaron Kusmec


amkusmec/FarmCPUpp documentation built on Dec. 25, 2021, 10:05 p.m.