|
Type |
Poster Presentation |
Area |
Physical Chemistry |
Room No. |
Event Hall |
Time |
4월 20일 (금요일) 11:00~12:30 |
Code |
PHYS.P-191 |
Subject |
Predicting the Solvation Free Energy of Alzheimer's Disease-induced Proteins Using Deep Learning |
Authors |
Jungeun Kim, Sihyun Ham* Department of Chemistry, Sookmyung Women's University, Korea |
Abstract |
Alzheimer's disease is known to be associated with the aggregation of Aβ42 and Tau43 proteins which are intrinsically disordered proteins (IDPs). Solvation free energies and internal energies of proteins reflect the free energy landscape of the proteins in water, and solvation free energy is also a key parameter to understand protein aggregation. In this study, we utilized the deep learning technique to develop machines predicting the solvation free energy of the proteins. We compared two machines using deep neural network (DNN) and convolution neural network (CNN), in which the contact maps of proteins are used directly or pre-processed by convolution. We report that the correlation coefficients between predicted and reference values of all the energies are greater than 0.90, and DNN predicts slightly better than CNN. This study also shows that protein contact maps are important input features for predicting solvation free energy. We believe that our result is useful in studying the thermodynamic properties of the proteins. |
E-mail |
coyoi@naver.com |
|