===================== INSTRUCTIONS ==============================

This document contains instructions for installing and using the LBLGXE R package.

Note: LBLGXE_1.1 is a new version implementing a faster algorithm. LBLGXE_1.0
is still available for download for now.

1. Install

For Linux users, download the compiled R package file: LBLGXE_1.1.tar.gz

For Windows users, download the compiled R package file for Windows:
LBLGXE_1.1.zip

Use the following function in R to install the package

>install.packages(pkgs="filename")
where "filename" should be replaced with the corresponding file name (LBLGXE_1.1.tar.gz) 
with the correct path or using setwd to get to the directory that contains the downloaded 
file before installing the package.

Alternatively, at the Linux command line, you may type 
"R CMD INSTALL PACKAGE_DIRECTORY/LBLGXE_1.1.tar.gz" 
where PACKAGE_DIRECTORY is the folder containing LBLGXE_1.1.tar.gz.


2. Usage

Note: Two other R packages, hapassoc and dummies, are required for running
LBL. You need to download these packages from CRAN before loading them as
shown below.

#Load the packages into R
> library(LBLGXE)
> library(hapassoc)
> library(dummies)

#Read the details about the package
>?LBLGXE

#Rad the details about the function and an example for runing the function.
>?LBL 

3. Example

Using the example data file: LBL.ex

#load the example dataset that is included in the package
> data(LBL.ex)

#You are now ready to run LBL to make inference on association of haplotypes and 
environmental covariates with the disease.
#Note: The defaul setting are burn.in = 10000, num.it = 50000 in the LBL function. 
However, for a quick check to make sure that the package is loaded properly, 
the example uses much smaller numbers.

> out.LBL<-LBL(LBL.ex, numSNPs=5, burn.in=100, num.it=1000)

#Note the results indicate that one rare haplotype (h11011) and one common 
haplotype (h11111), smoking, h11011 X smoking (rare haplotype - smoking 
interaction) are associated with the disease (CI of OR deosn't include and 
BF is large).


> freq=c(round(out.LBL$freq[-length(out.LBL$freq)],4),rep("-",length(out.LBL$freq)))
> names(freq)=names(out.LBL$OR[-1])
> ans <-cbind(freq,out.LBL$OR[-1],out.LBL$CI.OR[-1,2:3],out.LBL$BF[-1])
> names(ans) <- c("Freq", "OR", "Lower", "Upper", "BF")
> ans
               Freq        OR     Lower    Upper                 BF
h01100       0.2989 1.0541795 0.8849394 1.259554 0.0777079420938466
h10100       0.0048 0.7733712 0.2211603 2.207507  0.559935545827771
h11011       0.0064 3.1547735 1.3279157 7.319831               >100
h11100       0.1688 1.0258106 0.8276207 1.244102 0.0661298204003316
h11111       0.0967 1.8295711 1.4131744 2.483620               >100
h01100smoke1      - 1.0412932 0.8294621 1.260237 0.0824179535038615
h10100smoke1      - 1.2169130 0.4289920 4.628844  0.650352578557744
h11011smoke1      - 1.8738140 1.0083842 3.816533   1.86189649506451
h11100smoke1      - 0.9814284 0.7826987 1.317883 0.0712350561491991
h11111smoke1      - 0.9874874 0.7295273 1.406077 0.0934363964856611
smoke1            - 1.6936493 1.2473958 2.202809   94.3011238908769


4. References

Biswas S, Xia S, Lin, S. (2013) Detecting Rare Haplotype-Environment Interaction
with Logistic Bayesian LASSO. Genetic Epidemiology, in press.

Biswas S, and Lin, S. (2011) Logistic Bayesian Lasso for identifying association 
with rare haplotypes and application to age-related macular degeneration. 
Biometrics 68:587-597.

-----------------------------------------------------------------
If you have any questions, please email Shuang Xia at xia64.osu.edu, Swati Biswas 
at Swati.Biswas@unthsc.edu or Shili Lin at shili@stat.osu.edu