🍶Systems-biology approach to GWAS.
Switch branches/tags
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
include/gin
lib
src
test
.gitignore
.gitmodules
.travis.yml
CMakeLists.txt
LICENSE
README.md
install_gin.sh

README.md

gin

DOI Build Status

gin (GWAS Incorporating Networks) is a software framework aimed at improving biomarker discovery on genotyping data using a priori information, namely networks. It is the successor of SConES, the network guided multi-locus mapping method. It includes two executables (the original scones and shake, its extended version) as well as the gin library, ready to be used by other software, like martini.

Installation

gin requires CMake >= 3.2 to compile. To install, simply do

git clone --recursive https://github.com/hclimente/gin.git
gin/install_gin.sh

This will install gin, scones and shake in gin/build. If you prefer another installation path, add it is as first argument eg gin/install_gin.sh /usr/local.

Usage

You can analyze your GWAS data from the command line executables. If you wish to use R for your analysis, please refer to the R interface martini. The files used in these examples are available in test/data/case1.

Shake

This command is equivalent to running SConES:

shake --ped genotype --pheno phenotype.txt --net network.txt --depth 1

(See all available commands with shake --help.)

SConES

This is an example of how to run SConES:

scones genotype phenotype.txt network.txt 0.05 . additive 0

The arguments are (in order):

  • The prefix of the PED/MAP files, the phenotype and the network files.
  • The minor allele frequency to filter.
  • The output directory.
  • The genetic model.
  • The number of main principal components to be removed.

Credits

gin is based on easyGWAS, a C/C++ framework for computing genome-wide association studies and meta-analysis developed by dominikgrimm. easyGWAS includes several standard methods for performing GWAS, such as linear regression, logistic regression and popular linear mixed models (EMMAX, FaSTLMM) to also account for population stratification.