Home   Research   Resource   About us   Sitemap   
R Packages
 Funmap Package
 EIP Package
 fGWAS Package
 HaruPDF Package
Software
 3FunMap Software
 fGWAS Software

Home>> Resource >> Software >> fGWAS

Glasso model: Bayesian Group LASSO for the fGWAS software



Description:

The Glasso model is a command line program which operates from the command line or from a shell. After you finish the installation, you can find the root of fGWAS application folder. The Glasso program is in the "glasso" subfolder. It provides 4 commands, including SIMULATE, PLINK, SIMPLE and MERGE. The following picture shows how to get the help message for the PLINK command in the command window.

In this page, all commands are explained and some examples are shown. For more details, please refer to the introduction document.



Commands:

0). Common options for all commands of the Glasso model

The Glasso model has some common options for all commands. The following table lists all commonn options. Among these options, -par, -pcf, -log and -rdata just need to specify a filename, the end-users don't provide the contents for these files. The option of -cfg can be generated by the the GUI program of the fGWAS.

NoOptionsDescription
1-h or --help Print short help message and exit.
2-v or --version Print version info and exit.
3-d or --debug Turn on build-in debugging.
4-cfg=<in-file> Specify a system configuration file.
Here is an example.
5-pcf=<in-file> Specify a filename for process communication between the command process and the UI process.
6-res=<out-file> Specify a filename for the result dataset in user-defined format, output file.
7-log=<out-file> Specify a text filename for the log file, output file.
8-rdata=<out-file> Specify a RDATA filename for the result dataset, output file.

1). Options for the SIMULATE command

The command provides a simulation function to generate simulated data and do group lasso analysis.

NoOptionsDescription
1-g=<num> Indicate the group(chromosome) number for the result.
2-par=<in-file>Specify a parameter file for the simulation.
Here is an example
3-run=<0|9>Data generating(9) and analysis procedure (0) are available. The ananlysis includes variable selection and refit procedure. Defaultly simulation, variable selection and refit procedure are performed successive.
4-znorm=<0|1>Indicate whether the data of Z columns are normalized.
5-pheout=<out-file>Specify a CSV filename for the phenotype dataset, output file.
6-snpout=<out-file>Specify a CSV filename for the SNP dataset, output file.

Example

The following example shows how to generate a simulated dataset. The option of "-run=9" means to generate data only.

x:\> glasso SIMULATE -g=01 -par=test-par.par -run=9 -pheout=test-par.csv -snpout=test-par
...
SUMMARY(Options)
....
SUMMARY(Parameters)
....
Simulation is done successfully.

2). Options for the PLINK command

The command provides a function to load PLINK dataset and do group lasso analysis.

NoOptionsDescription
1-g=<num> (Required)Indicate the group (chromosome) number for the result.
2-tped=<in-file> Specify a TPED filename exported by PLINK software.
3-tfam=<in-file> Specify a TFAM filename exported by PLINK software.
4-phe=<in-file> Specify a CSV filename for phenotypic data.
5-znorm=<0|1> Indicate whether the data of Z columns are normalized.
6-run=<0|1|2> Variable selection (1), refit (2) procedure or both of them (0). Defaultly the refit will be performed after the variable selection is successful finished.
7-vsret=<in-file> If the refit will be performed alone, the result of variable selection must be provided. This file is produced by the commamd of '-run=0'.

Example

The following example shows how to load PLINK dataset. The option of "-run=1" means to do variable selection only.

x:\> glasso PLINK -g=01 -tped=../chr2.tped -tfam=../chr2.tfam -phe=../bmi.csv -run=1

3). Options for the SIMPLE command

The command provides a function to load SIMPLE dataset and do group lasso analysis. The SIMPLE data format is defined by the fGWAS software. The SIMULATE command can export genotypical data as the SIMPLE format.

NoOptionsDescription
1-g=<num> Indicate the group (chromosome) number for the result.
2-snp=<in-file> Specify a SIMPLE snp file, it can be simulated by the SIMULATE command.
Here is an example
3-phe=<in-file> Specify a CSV filename for phenotypic data.
Here is an example
4-znorm=<0|1> Indicate whether the data of Z columns are normalized.
5-run=<0|1|2> Variable selection (1), refit (2) procedureor both of them (0). Defaultly the refit will be performed after the variable selection is successful finished.
6-vsret=<in-file> If the refit will be performed alone, the result of variable selection must be provided. This file is produced by the commamd of '-run=0'.

Example

The following example shows how to load SIMPLE dataset. The option of "-run=0" means to do variable selection and refit procedure successively.

x:\> glasso SIMPLE -g=01 -snp=test-par.snp -phe=test-par.csv -run=0

4). Options for the MERGE command:

The command merges multiple results from the variable selection process and then do a refit process to get the final results for whole anlysis.

NoOptionsDescription
1-g=<num>Indicate the group(chromosome) number for the result.
2-mif=<in_file>Specify a mergence information file including the phenotypical file and multiple genotype file and correspoing results file.
Here is an example


Copyright© 2009-, Center for Statistical Genetics, Penn State University, All Rights Reserved.