abs

학술발표회초록보기

초록문의 abstract@kcsnet.or.kr

결제문의 member@kcsnet.or.kr

현재 가능한 작업은 아래와 같습니다.
  • 09월 10일 16시 이후 : 초록수정 불가능, 일정확인 및 검색만 가능

제124회 대한화학회 학술발표회, 총회 및 기기전시회 안내 Predicting protein-ligand binding affinity using the ensemble of 3D-convolutional neural networks

등록일
2019년 8월 21일 14시 53분 53초
접수번호
0715
발표코드
PHYS.P-159 이곳을 클릭하시면 발표코드에 대한 설명을 보실 수 있습니다.
발표시간
10월 17일 (목요일) 11:00~12:30
발표형식
포스터
발표분야
Physical Chemistry
저자 및
공동저자
Yongbeom Kwon, Juyong Lee*
Department of Chemistry, Kangwon National University, Korea
Accurate prediction of the binding affinity of a protein-ligand complex is essential for efficient and successful rational drug design. In this work, a new neural network model that predicts the binding affinity of a protein-ligand complex structure is developed. Our new model predicts the binding affinity of a complex using the ensemble of multiple independently trained networks that consist of multiple channels of 3D convolutional neural network[1] layers. Our model was trained using the 3740 protein-ligand complexes from the refined set of the PDBbind[2] database and tested using the 270 complexes from the core set. The benchmark results show that the correlation coefficient between the predicted binding affinities by our model and the experimental data is higher than 0.72, which is comparable with the state-of-the-art binding affinity prediction methods.[3] In addition, our method also ranks the relative binding affinities of possible multiple binders of a protein quite accurately. Last, we measured which structural information is critical for predicting binding affinity.

References
[1] LeCun, Y. et al., Nature, 521 (2015), 436–444.
[2] Liu, Z. et al., Acc. Chem. Res., 50 (2017), 302–309.
[3] Su, M. et al., J. Chem. Inf. Model., 59 (2019), 895–913.

상단으로