[Genabel-commits] r1138 - tutorials/GenABEL_general

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

Author: lckarssen
Date: 2013-03-13 20:56:54 +0100 (Wed, 13 Mar 2013)
New Revision: 1138

Modified:
   tutorials/GenABEL_general/ImputedDataAnalysis.Rnw
Log:
- A few small spelling fixes
- Workaround for ProbABEL bug #2575 (weird output file names)


Modified: tutorials/GenABEL_general/ImputedDataAnalysis.Rnw
===================================================================
--- tutorials/GenABEL_general/ImputedDataAnalysis.Rnw	2013-03-13 17:35:20 UTC (rev 1137)
+++ tutorials/GenABEL_general/ImputedDataAnalysis.Rnw	2013-03-13 19:56:54 UTC (rev 1138)
@@ -33,9 +33,9 @@
 
 Let us start with analysis of directly typed SNPs. For that, 
 we will use \texttt{mlreg} function of \GA{}. This function 
-implements ML-regression and Wald test of siginificance\footnote{
-In general the score test may be preferred becuse it is faster and 
-more robust
+implements ML-regression and Wald test of significance\footnote{
+In general the score test may be preferred because it is faster and 
+more robust.
 }. 
 This will later on allow us direct comparison with the results of \PA{}, 
 which implements the same testing procedure.
@@ -79,19 +79,19 @@
 @
 
 \begin{Exercise}
-It is known that rare variation in presence of 
+It is known that rare variation in the presence of 
 outliers can generate spurious associations. 
-Do you believe this is true association in this 
+Do you believe this is a true association in this 
 particular case? What you can do to check whether 
-this is true association or not?
+this is a true association or not?
 \end{Exercise}
 \begin{Answer}
 Firstly, you can check (by producing a cross-plot 
-of genotype vs. phenotype) if association is indeed 
+of genotype vs.~phenotype) if association is indeed 
 due to extreme phenotypic outliers. A related question is 
 whether the distribution is skewed. Additionally, 
 a permutation-based test can help establishing 
-correct p-value, taking into account the nature 
+correct $p$-value, taking into account the nature 
 of the data in question.  
  
 However, to give an ultimate answer, a replication 
@@ -102,7 +102,7 @@
 
 \section{Analysis of imputed data with \PA{}}
 
-Here, you will analyse imputed data. In \texttt{RData} 
+Here, you will analyse imputed data. In the \texttt{RData} 
 directory, you will find the necessary files: 
 \texttt{mach1.mldose.fvi} and \texttt{mach1.mldose.fvd} 
 (these files represent \texttt{mldose} data produced by 
@@ -113,10 +113,10 @@
 We will start with producing a phenotypic data file for the 
 use with \PA{}:
 <<>>=
-write.table(data.frame(id=idnames(df500),rcT=rcT),
-       file="rcT.PHE",quote=F,row.names=F)
+write.table(data.frame(id=idnames(df500), rcT=rcT),
+       file="rcT.PHE", quote=FALSE, row.names=FALSE)
 @
-next, try command '\texttt{system("head rcT.PHE")}' to check few first 
+next, try the command '\texttt{system("head rcT.PHE")}' to check the few first 
 lines of the file.
 
 At this moment, leave \texttt{R} (or, rather, start new console!), 
@@ -129,22 +129,22 @@
 yourname at server> palinear --pheno rcT.PHE --info mach1.out.mlinfo /
                  --dose mach1.mldose.fvi
 \end{verbatim}
-Do not forget to check that you start analysis in right directory, i.e. 
+Do not forget to check that you start the analysis in right directory, i.e. 
 all files (\texttt{rcT.PHE}, \texttt{mach1.out.mlinfo}, 
 \texttt{mach1.mldose.fvi}, \texttt{mach1.mldose.fvd}) 
-are present in the working directory (use command '\texttt{ls}' 
+are present in the working directory (use the '\texttt{ls}' command 
 from the console). 
 
 <<echo=false,hide=true>>=
 system("cp RData/mach1* .")
 system("cp RData/rcT.PHE .")
-palCmd <- "palinear --pheno rcT.PHE --info mach1.out.mlinfo --dose mach1.mldose.fvi"
+palCmd <- "palinear --pheno rcT.PHE --info mach1.out.mlinfo --dose mach1.mldose.fvi -o regression"
 system(palCmd)
 @
 
 Now, you can return to \texttt{R} and load the analysis results:
 <<>>=
-qtsPal <- read.table("regression_add.out.txt",head=T,strings=F)
+qtsPal <- read.table("regression_add.out.txt", head=T, strings=F)
 qtsPal[1:5,]
 @
 As you see, there is not $P$-value produced in the \PA{} output, and