Inspired by our recently developed FarmCPU method (PLoS Genetics) and its R package, we developed a new method, named BLINK (Bayesian-information and Linkage-disequilibrium Iteratively Nested Keyway). The new method releases the requirement that causative genes are evenly distributed on genome and consequently boosts statistical power. The REstricted Maximum Likelihood (REML) by using random effect model in FarmCPU is also replaced with Bayesian Information Content (BIC) by using fixed effect model to further improve computing speed. The manuscript on BLINK algorithms and performances is available at bioRxiv . Two packages were available at GitHub: C version and R version. Both packages contain user manuals and demo data. The R package contains R source code for free usage. The C package contains executable program that is also free to analyze datasets under 50 million data points (number of individuals x number of markers). Releases of limitation on the data points, or Licenses for BLINK C source code are available at the Office of Commercialization at Washington State University. Although the C version and the R version have the same statistical power, the C version of BLINK is not only superior to FarmCPU, but also to R version of BLINK. A dataset with half million markers and half million individuals can be analyzed within an hour with C version of BLINK, compared with days by using FarmCPU. Questions and comments can be addressed to Dr. Meng Huang (Email: for C version, or Dr. Yao Zhao (Email: for R version.