A feedback generative model for De Novo drug design = 약물 디자인을 위한 피드백 생성 모델

Cited 0 time in webofscience Cited 0 time in scopus
  • Hit : 33
  • Download : 0
DC FieldValueLanguage
dc.contributor.advisorKim, Dong Sup-
dc.contributor.advisor김동섭-
dc.contributor.authorHan, ShangJin-
dc.date.accessioned2021-05-13T19:37:14Z-
dc.date.available2021-05-13T19:37:14Z-
dc.date.issued2020-
dc.identifier.urihttp://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=925096&flag=dissertationen_US
dc.identifier.urihttp://hdl.handle.net/10203/284937-
dc.description학위논문(석사) - 한국과학기술원 : 바이오및뇌공학과, 2020.8,[iii, 32 p. :]-
dc.description.abstractArtificial intelligence technology has made significant progress in many fields of medicine. It can integrate a large amount of genetic and molecular data, pharmacological data in the field of new drug research and development, conduct effective target screening and drug design, save drug research and development costs, and shorten drug development time. This article takes Discoidin Domain Receptor 1 (DDR1) as the target and generates potential DDR1 kinase inhibitors, which are used for treatment of fibrosis and other diseases. This article uses a modified generative tensorial reinforcement learning (GENTRL) as the core of the model, uses Self-Referencing Embedded Strings (SELFIES) instead of the most commonly used Simplified Molecular Input Line Entry System (SMILES) as the molecular encoding method, and uses a feedback mechanism and optimize model performance and generate high-quality potential DDR1 kinase inhibitors. We show that the modified GENTRL model can effectively generate highly effective molecules, and SELFIES can improve the valid percentage of generated molecules compared to SMILES and allow the model to produce highly valid molecules in a shorter time. In addition, the feedback mechanism can effectively increase the docking score of the generated molecules with target protein, generating compounds with better binding affinity to the DDR1 protein site.-
dc.languageeng-
dc.publisher한국과학기술원-
dc.subjectdeep generative model-
dc.subjectDDR1 kinase inhibitors-
dc.subjectfeedback-
dc.subjectvariational autoencoder-
dc.subjectreinforcement learning-
dc.subject깊은 생성 모델-
dc.subject피드백-
dc.subject강화 학습-
dc.titleA feedback generative model for De Novo drug design = 약물 디자인을 위한 피드백 생성 모델-
dc.typeThesis(Master)-
dc.identifier.CNRN325007-
dc.description.department한국과학기술원 :바이오및뇌공학과,-
dc.contributor.alternativeauthorShangJin Han-
Appears in Collection
BiS-Theses_Master(석사논문)
Files in This Item
There are no files associated with this item.

qr_code

  • mendeley

    citeulike


rss_1.0 rss_2.0 atom_1.0