Active responses from experts play an essential role in the knowledge discovery of SAR (structure activity relationships) from drug data. Experts often think of hypotheses, and they want to reflect these ideas to the attribute generation and selection processes. Authors have analyzed SAR of dopamine antagonists using the cascade model. In this paper, we generated attributes indicating the presence of hydrogen-bonded fragments from 3D coordinates of molecules, which were suggested by experts. The selection of attributes by experts has been shown to be useful in obtaining rules that lead to valuable knowledge.