Improving the computation efficiency of polygenic risk score modeling: faster in Julia