In defense of genetic association studies