MVP: MVP, Memory-efficient, Visualization-enhanced,...

In XiaoleiLiuBio/MVP: Memory-Efficient, Visualize-Enhanced, Parallel-Accelerated GWAS Tool

MVP, Memory-efficient, Visualization-enhanced, Parallel-accelerated

Description

Object 1: To perform GWAS using General Linear Model (GLM), Mixed Linear Model (MLM), and FarmCPU model Object 2: To calculate kinship among individuals using Varaden method Object 3: Estimate variance components using EMMA, FaST-LMM, and HE regression Object 4: Generate high-quality figures

Usage

MVP(
  phe,
  geno,
  map,
  K = NULL,
  nPC.GLM = NULL,
  nPC.MLM = NULL,
  nPC.FarmCPU = NULL,
  CV.GLM = NULL,
  CV.MLM = NULL,
  CV.FarmCPU = NULL,
  REML = NULL,
  maxLine = 10000,
  ncpus = detectCores(logical = FALSE),
  vc.method = c("BRENT", "EMMA", "HE"),
  method = c("GLM", "MLM", "FarmCPU"),
  maf = NULL,
  p.threshold = NA,
  QTN.threshold = 0.01,
  method.bin = "static",
  bin.size = c(5e+05, 5e+06, 5e+07),
  bin.selection = seq(10, 100, 10),
  maxLoop = 10,
  permutation.threshold = FALSE,
  permutation.rep = 100,
  memo = NULL,
  outpath = getwd(),
  col = c("#4197d8", "#f8c120", "#413496", "#495226", "#d60b6f", "#e66519", "#d581b7",
    "#83d3ad", "#7c162c", "#26755d"),
  file.output = TRUE,
  file.type = "jpg",
  dpi = 300,
  threshold = 0.05,
  verbose = TRUE
)

Arguments

phe	phenotype, n * 2 matrix, n is sample size
geno	Genotype in bigmatrix format; m * n, m is marker size, n is sample size
map	SNP map information, SNP name, Chr, Pos
K	Kinship, Covariance matrix(n * n) for random effects, must be positive semi-definite
nPC.GLM	number of PCs added as fixed effects in GLM
nPC.MLM	number of PCs added as fixed effects in MLM
nPC.FarmCPU	number of PCs added as fixed effects in FarmCPU
CV.GLM	covariates added in GLM
CV.MLM	covariates added in MLM
CV.FarmCPU	covariates added in FarmCPU
REML	a list contains ve and vg
maxLine	the number of markers handled at a time, smaller value would reduce the memory cost
ncpus	number of cpus used for parallel
vc.method	methods for estimating variance component("EMMA" or "HE" or "BRENT")
method	the GWAS model, "GLM", "MLM", and "FarmCPU", models can be selected simutaneously, i.e. c("GLM", "MLM", "FarmCPU")
maf	the threshold of minor allele frequency to filter SNPs in analysis
p.threshold	if all p values generated in the first iteration are bigger than p.threshold, FarmCPU stops
QTN.threshold	in second and later iterations, only SNPs with lower p-values than QTN.threshold have chances to be selected as pseudo QTNs
method.bin	'static' or 'FaST-LMM'
bin.size	window size in genome
bin.selection	a vector, how many windows selected
maxLoop	maximum number of iterations
permutation.threshold	if use a permutation cutoff or not (bonferroni cutoff)
permutation.rep	number of permutation replicates
memo	Character. A text marker on output files
col	for color of points in each chromosome on manhattan plot
file.output	whether to output files or not
file.type	figure formats, "jpg", "tiff"
dpi	resolution for output figures
threshold	a cutoff line on manhattan plot, 0.05/marker size
verbose	whether to print detail.
tmppath	the path of the temporary file

Details

Build date: Aug 30, 2017 Last update: Dec 14, 2018

Value

a m * 2 matrix, the first column is the SNP effect, the second column is the P values Output: MVP.return$map - SNP map information, SNP name, Chr, Pos Output: MVP.return$glm.results - p-values obtained by GLM method Output: MVP.return$mlm.results - p-values obtained by MLM method Output: MVP.return$farmcpu.results - p-values obtained by FarmCPU method

Author(s)

Lilin Yin, Haohao Zhang, and Xiaolei Liu

Examples

phePath <- system.file("extdata", "07_other", "mvp.phe", package = "rMVP")
phenotype <- read.table(phePath, header=TRUE)
print(dim(phenotype))
genoPath <- system.file("extdata", "06_mvp-impute", "mvp.imp.geno.desc", package = "rMVP")
genotype <- attach.big.matrix(genoPath)
print(dim(genotype))
mapPath <- system.file("extdata", "06_mvp-impute", "mvp.imp.geno.map", package = "rMVP")
map <- read.table(mapPath , head = TRUE)

opts <- options(rMVP.OutputLog2File = FALSE)

mvp <- MVP(phe=phenotype, geno=genotype, map=map, maxLoop=3,
  method=c("GLM", "MLM", "FarmCPU"), file.output=FALSE, ncpus=1)
str(mvp)

options(opts)

XiaoleiLiuBio/MVP documentation built on Sept. 27, 2024, 7:44 a.m.