2014 年 13 巻 6 号 p. 303-304
hA2AAR antagonists are developed as drugs for the treatment of neurodegenerative disorders. First, we generated the developed structures from a 1,2,4-triazole derivative by using transformation rules of StarDrop/Nova and BIOSTER. In one case, we got 73,452 compounds of the first and second generations by using all transformation rules. There were 4,069 compounds with more suitable ADME than the first compound. In another case, we generated structures of the first, second and third generations, selecting ADME score's top 300 compounds each generation. As a result, we got the compounds having the higher score of ADME with higher probability when we used all transformation rules. Next, we evaluated 325 compounds with suitable ADME by using Torch3D. Torch3D can compare the 3D compound with 2D compounds based on electrostatic/hydrophobic/aromatic/steric characters. Finally, we got the best compound with not only the higher score of ADME than the first one but also suitable 3D molecular characters for docking to hA2AAR.