Collectively, we have presented practical tips that emphasize cutting-edge insights and describe evolving professional standards. In addition, a number of our points are focused on safeguarding against the risks inherent to data science and deep learning. These risks include the over- and mis-interpretation of models, poor generalizability, and the potential to harm others. However, we want to strongly emphasize that when used ethically and responsibly, deep learning techniques have the potential to add tremendous value across a diverse range of contexts. After all, these techniques have already shown a remarkable capacity to meet or exceed the performance of both humans and traditional algorithms and have the potential to uncover biomedical insights that drive discovery and innovation. By taking a comprehensive and careful approach to deep learning based on critical thinking about research questions, planning to maintain rigor, and discerning how work might have far-reaching consequences with ethical dimensions, the life science community can advance reproducible, interpretable, and high-quality science that is enriching and beneficial for both scientists and society.