The role of different sampling methods in improving biological activity prediction using deep belief network

• Published in: Journal of computational chemistry Vol. 38; no. 4; pp. 195 - 203
• Main Authors: Ghasemi, Fahimeh, Fassihi, Afshin, Pérez‐Sánchez, Horacio, Mehri Dehnavi, Alireza
• Format: Journal Article
• Language: English
• Published: United States Wiley Subscription Services, Inc 05.02.2017
• Subjects: Algorithms; Biological activity; biological activity prediction; Computer Simulation; deep belief network; deep neural network; Drug Design; initialization; Mathematical problems; Models, Theoretical; Molecular chemistry; Neural networks; Neural Networks (Computer); Parameter estimation; Quantitative Structure-Activity Relationship; Random number sampling; Statistical methods; statistical modeling; biological activity prediction; deep belief network; initialization; statistical modeling; deep neural network
• ISSN: 0192-8651; 1096-987X; 1096-987X
• DOI: 10.1002/jcc.24671

Thousands of molecules and descriptors are available for a medicinal chemist thanks to the technological advancements in different branches of chemistry. This fact as well as the correlation between them has raised new problems in quantitative structure activity relationship studies. Proper parameter initialization in statistical modeling has merged as another challenge in recent years. Random selection of parameters leads to poor performance of deep neural network (DNN). In this research, deep belief network (DBN) was applied to initialize DNNs. DBN is composed of some stacks of restricted Boltzmann machine, an energy‐based method that requires computing log likelihood gradient for all samples. Three different sampling approaches were suggested to solve this gradient. In this respect, the impact of DBN was applied based on the different sampling approaches mentioned above to initialize the DNN architecture in predicting biological activity of all fifteen Kaggle targets that contain more than 70k molecules. The same as other fields of processing research, the outputs of these models demonstrated significant superiority to that of DNN with random parameters. © 2016 Wiley Periodicals, Inc. High throughput virtual screening is a kind of computational approach in drug discovery challenged many fundamental problems such as prone to over‐fitting. The novel solution to avoid them is pre‐training network parameters by deep belief network (DBN). The main problem in applying DBN is calculation of log likelihood gradient. Different sampling approaches have been suggested. The results of this study demonstrated that DBN could improve the ability of DNN to provide high quality predicting models.