Recent advances of medical informative tools and high throughput technologies have made large scale data routinely accessible to researchers, administrators, and policy-makers. Large scale data provides unprecedented opportunities for new and innovative approaches to science and medical research. On the other hand, this ``data deluge'' also poses new challenges and critical barriers for biostatisticians and scientists as existing statistical methods are rendered unfeasible for analyzing these large scale datasets. In this talk I will discuss a new suite of scalable sparse regression tools for large scale medical data. Specifically, we will focus on two common types of large scale data: 1) high dimensional (number of predictors in thousands, or tens of thousands), massive sample-size (thousands or tens of thousands) data; and 2) high dimensional (number of predictors in thousands, tens of thousands), small sample-size (tens or hundreds) data.