분자구조 유사도를 활용한 약물 효능 예측 알고리즘 연구

• Published in: Journal of biomedical engineering research Vol. 43; no. 4; pp. 230 - 240
• Main Authors: 정화영, 송창현, 조혜연, 기재홍, Jeong, Hwayoung, Song, Changhyeon, Cho, Hyeyoun, Key, Jaehong
• Format: Journal Article
• Language: Korean
• Published: 대한의용생체공학회 01.08.2022
• Subjects: 의공학; Drug similarity; SMILES; ATC code; Drug regeneration; LINGO similarity
• ISSN: 1229-0807; 2288-9396

Drug regeneration technology is an efficient strategy than the existing new drug development process, which requires large costs and time by using drugs that have already been proven safe. In this study, we recognize the importance of the new drug regeneration aspect of new drug development and research in predicting functional similarities through the basic molecular structure that forms drugs. We test four string-based algorithms by using SMILES data and searching for their similarities. And by using the ATC codes, pair them with functional similarities, which we compare and validate to select the optimal model. We confirmed that the higher the molecular structure similarity, the higher the ATC code matching rate. We suggest the possibility of additional potency of random drugs, which can be predicted through data that give information on drugs with high molecular similarities. This model has the advantage of being a great combination with additional data, so we look forward to using this model in future research.