PGC: Polynomial genomic control

In GenABEL: genome-wide SNP association analysis

Description

This function estimates the genomic controls for different models and degrees of freedom, using polinomial function. Polinomial coefficients are estimated by optimizing different error functions: regress, median, ks.test or group regress.

Usage

  PGC(data, method = "group_regress", p, df, pol.d = 3,
    plot = TRUE, index.filter = NULL, start.corr = FALSE,
    proportion = 1, n_quiantile = 5, title_name = "Lambda",
    type_of_plot = "plot", lmax = NULL, color = "red")

Arguments

data	Input vector of Chi square statistic
method	Function of error to be optimized. Can be "regress", "median", "ks.test" or "group_regress"
p	Input vector of allele frequencies
df	Number of degrees of freedom
pol.d	The degree of polinomial function
plot	If TRUE, plot of lambda will be produced
start.corr	For regress method use it only when you want to make calculations faster
index.filter	Index of variables in data vector, that will be used in analysis if zero - all variables will be used
proportion	The proportion of lowest P (Chi2) to be used when estimating the inflation factor Lambda for "regress" method only
n_quiantile	The number of groups for "group_regress" method
title_name	The title name for plot
type_of_plot	For developers only
lmax	The threshold for lambda for plotting (optional)
color	The color of the plot

Value

A list with elements

data	Output vector corrected Chi square statistic
b	Polinomial coefficients

Author(s)

Yakov Tsepilov

Examples

require(GenABEL.data)
data(ge03d2)
ge03d2 <- ge03d2[seq(from=1,to=nids(ge03d2),by=2),seq(from=1,to=nsnps(ge03d2),by=3)]
qts <- mlreg(dm2~1,data=ge03d2,gtmode = "additive")
chi2.1df <- results(qts)$chi2.1df
s <- summary(ge03d2)
freq <- s$Q.2
result=PGC(data=chi2.1df,method="median",p=freq,df=1, pol.d=2, plot=TRUE, lmax=1.1,start.corr=FALSE)

GenABEL documentation built on May 30, 2017, 3:36 a.m.