plotMultiQTL: Plot QTL point estimates and confidence intervals as...

In jtlovell/multiQTL: Functions to extract and plot statistics from multiple QTL modesl

Description

plotMultiQTL provides a system to visualize QTL distributions for many traits. This function integrates directly with R/qtl and the stepwiseStats functions.

Usage

plotMultiQTL(cross, stats = NULL, phes = NULL, chrs = NULL, peak = NULL, right = NULL, left = NULL, col = NULL, chr.subset = NULL, ylabelcex = NULL, rugsize = NULL, cex = NULL, pch = 19, lty = 1, lwd = 1, plotQTLdensity = TRUE, binwidth = 1, adj.ylabsize = TRUE, colbychr = TRUE, palette = rainbow, showConfidenceInterval = TRUE, showPointEstimate = TRUE, outline = FALSE, background = TRUE, plotNullPheno = FALSE, setmargin = NULL, ...)

Arguments

cross	R/qtl cross object, required
stats	Output from stepwiseStats, or a dataframe with column names that match: "phenotype","chromosome","position","lowCIpos","hiCIpos"
phes	If stats is NULL, required. A character vector of phenotype names.
chrs	If stats is NULL, required. A character vector of chromosome names/numbers.
peak	If stats is NULL and showPointEstimate=TRUE, required. A numeric vector of QTL peak positions.
right	If stats is NULL and showConfidenceInterval=TRUE, required. A numeric vector of right confidence interval bounds names.
left	If stats is NULL and showConfidenceInterval=TRUE, required. A numeric vector of right confidence interval bounds names.
col	A vector of line and point color, indexed by the line in stats, or position in vectors. Can be a single value or a vector with a length identical to the number of rows in stats, or the length of the data Overridden by colbychr
chr.subset	Optional. Numeric/character vector of chromosomes by which to subset
ylabelcex	the y axis label size. Can be a single value or a vector with a length identical to the number of rows in stats, or the length of the data
rugsize	The height of the genetic map segments. If rugsize=0, suppress the genetic map.
cex	The size of the points representing the point estimates. Can be a single value or a vector with a length identical to the number of rows in stats, or the length of the data
pch	The shape of the points representing the point estimates. Can be a single value or a vector with a length identical to the number of rows in stats, or the length of the data.
lty	The line style for lines representing the confidence intervals. Can be a single value or a vector with a length identical to the number of rows in stats, or the length of the data
lwd	The linewidth for the lines representing the confidence intervals. Can be a single value or a vector with a length identical to the number of rows in stats, or the length of the data
plotQTLdensity	Logical. Should the function plot a density distribution of QTLs at the top of the figure. See details.
binwidth	Numeric. If plotQTLdensity=T, this specifies the binwidth (bw) argument in density in terms of cM. See details.
adj.ylabsize	Logical. Should the function choose the best y axis label cex
colbychr	Logical. Should the QTL be colored by chromosomes
palette	If colbychr=T, specifies the palette to use to color chromosomes and points
showConfidenceInterval	Logical. Should the confidence intervals be plotted?
showPointEstimate	Logical. Should the point estimates be plotted?
outline	Logical. Should a box be placed around the plotting area?
background	Logical, should a transparent background be placed that helps to visually separate the chromosomes?
plotNullPheno	Logical. Should phenotypes without QTL on the subsetted chromosomes be retained? If false, only phenotypes with QTL are plotted
setmargin	Optional. The par margin vector of 4 margin sizes.
...	Additional arguments passed on to plot. Specify xlab to "chromosome" to reproduce R/qtl - style scanone x axes.

Details

Plots points and segments using R base functions. The plotQTLdensity argument invokes the R base function density which calculates kernal density of QTL from the lowest cM position of a QTL on the first chromosome to the highest cM position on the last chromosome. Specifying binwidth changes the "bw" argument within density. Higher numbers cause greater smoothing of the profile.

Value

A plot of QTL positions.

Author(s)

John T. Lovell

References

Lovell et al. (2015) Exploiting differential gene expression and epistasis to discover candidate genes for drought-associated QTLs in Arabidopsis thaliana. The Plant Cell: Vol. 27: 969<e2><80><93>983

Examples

library(qtl)
library(plyr)
#use the multitrait dataset first
data(multitrait)
cross <- multitrait
qtlphes<-phenames(cross)[1:8]
cross <- calc.genoprob(cross, step=2.5)
modelList<-lapply(qtlphes, function(x) {
  stepwiseqtl(cross, penalties=c(3,4,3),max.qtl=3, pheno.col=x, method="hk", keeptrace=TRUE, verbose=FALSE, keeplodprofile=TRUE)
})
names(modelList)<-qtlphes

stepParsed<-lapply(qtlphes, function(x){
  stepwiseStats(cross, model.in= modelList[[x]], phe=x, covar=NULL, ci.method="drop", drop=1.5, plot=FALSE, printout=TRUE)
})
statsDF<-ldply(stepParsed, data.frame)
plotMultiQTL(cross=cross, stats=statsDF, ylabelcex=.4, binwidth=5)

jtlovell/multiQTL documentation built on May 20, 2019, 3:14 a.m.