abs

학술발표회초록보기

초록문의 abstract@kcsnet.or.kr

결제문의 member@kcsnet.or.kr

현재 가능한 작업은 아래와 같습니다.
  • 09월 04일 17시 이후 : 초록수정 불가능, 일정확인 및 검색만 가능

제114회 대한화학회 학술발표회, 총회 및 기기전시회 안내 Quantitative Structure-Activity Relationship(QSAR) modelling for odor thresholds of odorous molecules

등록일
2014년 8월 28일 15시 57분 46초
접수번호
1171
발표코드
MEDI.P-972 이곳을 클릭하시면 발표코드에 대한 설명을 보실 수 있습니다.
발표시간
10월 15일 (수요일) 16:00~19:00
발표형식
포스터
발표분야
의약화학
저자 및
공동저자
서민, 이종혁, 이성광*
한남대학교 화학과, Korea
The odor thresholds are detection thresholds for odor substances. The detection of odors in humans occurs at the nasal cavity through the olfactory nerve. These odors cause negative consequences, such as discomfort or annoyance or sick house syndrome. Therefore, we can prevent that warning signals of gas leakage or similar accidents by prediction of odor thresholds. In this study, QSAR models for predicting odor thresholds of odorous molecules with 2D molecular descriptors have been developed. Odor thresholds value for 223 compounds were obtained Nagata data set. Then, the data set was splitted into a training set for 111 compounds and a test set for 112 compounds by sphere exclusion method(SEM). Effective subset of descriptors was selected on forward selection method with multiple linear regression (MLR) and support vector machine(SVM). All models were evaluated by bootstrap method and y-scrambling. The results indicate that new QSAR models aids in designing a library a new odorous compounds with better predicted response.

상단으로