[Genabel-commits] r674 - pkg/VariABEL/man

[noreply at r-forge.r-project.org]

Author: maksim
Date: 2011-03-07 15:14:29 +0100 (Mon, 07 Mar 2011)
New Revision: 674

Modified:
   pkg/VariABEL/man/var.test.homogeneity.Rd
Log:
help update

Modified: pkg/VariABEL/man/var.test.homogeneity.Rd
===================================================================
--- pkg/VariABEL/man/var.test.homogeneity.Rd	2011-03-07 10:59:16 UTC (rev 673)
+++ pkg/VariABEL/man/var.test.homogeneity.Rd	2011-03-07 14:14:29 UTC (rev 674)
@@ -1,39 +1,39 @@
 \name{var.test.homogeneity}
 \alias{var.test.homogeneity}
-\title{function to perfome variance homogeneity test of different genotypic groups}
+\title{Function performs compariosn of genotypic variances.}
 \description{
-function to perfome variance homogeneity test of different genotypic groups	
+Presence of significant difference of genotypic variances points to possible 
+		Significant variance difference points to possible presence of interaction between a tested SNP and a factor (or set of factors).
 }
 \usage{
-var.test.homogeneity(trait="", formula="", trait_df, gwa_data, dir, analysis_name_specific, how_many_sigma_drop=NULL, top_snp_num_figure=3, analysis_type = "AAvsABvsBB")
+var.test.gwaa(formula, genodata, phenodata, genodata_info=NULL, testname="sqlm", analysis_type="AAvsABvsBB")
 }
 \arguments{
-  \item{trait}{Name of trait you are analysing (data.frame trait_df has to contain a column with this name). If it absenses or equals to "" than the trait name must be specifed in "formula"}
-  \item{formula}{Object of a class "formula". It is used in case of any adljustment nessacarity. If it absenses or equals to "" than trait must be specifed}
-  \item{trait_df}{Object of class "data.frame". It contains columnes with analysing trait and covariates. Should contain column with id names.}
-  \item{gwa_data}{Object of class "gwaa.data"}
-  \item{dir}{Directory name where all results will be put. It should be created befored analysis starts.}
-  \item{analysis_name_specific}{Any set of characters. It figures as a part of a file names in output.}
-  \item{how_many_sigma_drop}{Cut off point to exclude outliers. Normally distributed trait is assumed.}
-  \item{top_snp_num_figure}{For how many SNPs to make boxplot.}
-  \item{analysis_type}{Which test is performed. Possible values: "AAvsABvsBB" (testing all three genotipc groups),
-		 	"AAvsABandBB" (testing AA against AB and BB together), "ABvsAAandBB", "BBvsAAandAB".}
+
+	\item{formula} Regression model used for analysis. In the first stage linear regression is run to exclude main snp effect. In this stage adjustment for covariates is performed.
+	\item{genodata} The genotypes data in format of genabel or databel object
+  \item{phenodata} The phenotypes data in format of data.frame object
+	\item{genodata_info} The file with snp information (name, position). Used if genodata is databel object.
+	\item{testname} Name of variance heterogeneity test to perform. sqlm (for imputed genotype data), levene, and bartlett test are supported. 
+	\item{analysis_type} Type of analsysis to perform. AAvsABvsBB - additive model where B allele additivly increase risk, AAvsABandBB - group AA tested agains AB and BB, ABvsAAandBB - AB against AA and BB, BBvsAAandAB - BB against AA and AB. Only available for typed snps.
 }
 \details{
-var.test.homogeneity function perform analisis of variance homogeneity for easch SNP in data "gwa_data". The main result is pval of variance homogeneity test for each SNP.
-Bartlett's test is currently used. This assumes normality in a trait distribution. Even a couple of outliers can spoil a whole pictures. 
-Input parametere "how_many_sigma_drop" should be used to protect overestimation of pval.
 
+The function var.test.gwaa tests for difference in genotypic variances. This difference points to presence of possible interaction between the tested SNP 
+and some factor. In the case sqlm test the analysis consists of two stage: firstly the regular GWA id done where regression analysis is performed with 
+covariates specified in the input parameter formula, in the second stage the regression analysis is performed with using residuals from the first stage and 
+a sno as a covariate.
 }
 
 \value{
-As result the file "lambda.doc" is created which contains inflation factor, file with test statistics (name is qt1__adj__sex_age_test_results.txt 
-				for example in case if you analyse "qt1" with adjustment on "sex" and "age"), file those contains id names which are excluded due to
-	 	normality resoans, manhettan plot, two qqplots (for chi2 amd pval) and boxplots for top SNPs.
+	The ouput is a data.frame object. The table contains the chisq of variance heterogeneity test (the name is chisq) the effects and standart errors of all covariates included into regression model, 
+	main snp effect (the names are snp_eff and snp_se). In the case of sqlm test the columns snp_eff_dispertion and snp_se_dispertion contain
+	effect of a snp on squared vallues of a trait.
+
 	}
   
 
-%\references{ ~put references to the literature/web site here ~ }
+%\references{ http://www.biomedcentral.com/1471-2156/11/92/abstract }
 \author{Maksim Struchalin}
 %\note{ ~~further notes~~ 
 %}
@@ -41,14 +41,13 @@
 	\code{\link{var.meta}},
 }
 \examples{
+	library(GenABEL)
 	data(srdta)
-	mytrait_df <- srdta at phdata[,c("id","qt1","sex","age")]
-	mydir <- "mydir_var_analysis"
-	dir.create(mydir)
-	var.test.homogeneity(formula=qt1 ~ sex + age, 
-					trait_df=mytrait_df, gwa_data=srdta, dir=mydir, analysis_name_specific="test", how_many_sigma_drop=3)
+	result1 <- var.test.gwaa(bt~qt1+qt2, genodata=srdta at gtdata, phenodata=srdta at phdata)
+	
+#if there is no covariates needed:
+	result2 <- var.test.gwaa("bt", genodata=srdta at gtdata, phenodata=srdta at phdata)
 
-
 }
 % Add one or more standard keywords, see file 'KEYWORDS' in the
 % R documentation directory.