BLINK: A Package for Next Level of Genome Wide Association Studies with Both Individuals and Markers in Millions
Inspired by our recently developed FarmCPU method, we continue our practice on modeling and software engineering. One of the outcomes is BLINK written in C by Meng Huang. Both real and simulated data analyses demonstrated that BLINK improves statistical power further and deliver a remarkable computing speed. Now, a dataset with half million markers and half million individuals can be analyzed within an hour, compared with days by using FarmCPU. The executable program of BLINK can be freely used to analyze datasets up to 50 million data points (number of individuals x number of markers) in size. Releases of limitation on the data points, or Licenses for BLINK source code are available at The Office of Commercialization at Washington State University (https://oc.store.wsu.edu/blink.html). Questions and comments can be made at BLINK forum (https://groups.google.com/forum/#!forum/blink). The executable program of BLINK, help documents, demonstration data, and tutorials are freely available here.
You can find the R version here: https://github.com/YaoZhou89/BLINK
OpenCL package instalation on Linux (For parallel computing)
OpenCL package isn't pre-installed on most Linux OS. Please download installation package, AMDAPPSDK-3.0.zip. Unzip this folder and use ./install.sh to install OpenCL. After successfully installed OpenCL, please put BLINK execuatble file and four kernel files into same folder.
If you just want to use single thread computing function of BLINK, please just download single thread version (blink_linux) and use it without any environment setting.
Sometimes, you may need to give the root right to BLINK executable file using command
chmod 777 blink_linux