Prof. Keiichi Inoue (visiting researcher) and Prof. Ichiro Takeuchi (team leader) of the Data-Driven Biomedical Science team at the AIP Center have published the results of their collaborative research on photobiology and machine learning in Communication Biology. The goal of this research project is to discover rhodopsin proteins that absorb light at high wavelengths from a large number of naturally occurring candidates. Using machine learning to predict absorption wavelengths, we screened 39 promising rhodopsin proteins from 3022 candidates, and experimentally measured the absorption wavelengths of the 39 screened rhodopsin proteins. As a result, 32 of the 39 proteins were found to absorb at higher wavelengths than the standard, as predicted. Conventionally, the discovery of such proteins required repeated trial-and-error by biologists, which was very costly. This study is one of the demonstrative examples of how machine learning predictions can be useful in facilitating scientific research.