Many algorithms have evolved in the past decade. Genetic analysis and deep learning are two representative groups of them. With the progress in the big data era and decreasing cost of data collection, these algorithms are being applied to a larger volume of data, leading to the new development of efficient systems in these domain areas. We evaluate the performance of some new emerging salable genetic analysis tools to provide practitioners with some guidelines. We also propose a new paradigm that we call intermittent human-in-the-loop model selection to mitigate pains in deep learning model selection.