CRNNTL: convolutional recurrent neural network and transfer learning for QSAR modelling
This work addresses QSAR modeling for drug discovery, offering an incremental improvement by integrating existing neural network techniques with transfer learning to handle small biological activity datasets.
The authors tackled QSAR modeling by proposing CRNNTL, a method combining convolutional and recurrent neural networks with transfer learning, and demonstrated its effectiveness on 20 benchmark datasets and in knowledge transfer for small datasets, achieving efficient performance in overcoming data scarcity.
In this study, we propose the convolutional recurrent neural network and transfer learning (CRNNTL) for QSAR modelling. The method was inspired by the applications of polyphonic sound detection and electrocardiogram classification. Our strategy takes advantages of both convolutional and recurrent neural networks for feature extraction, as well as the data augmentation method. Herein, CRNNTL is evaluated on 20 benchmark datasets in comparison with baseline methods. In addition, one isomers based dataset is used to elucidate its ability for both local and global feature extraction. Then, knowledge transfer performance of CRNNTL is tested, especially for small biological activity datasets. Finally, different latent representations from other type of AEs were used for versatility study of our model. The results show the effectiveness of CRNNTL using different latent representation. Moreover, efficient knowledge transfer is achieved to overcome data scarcity considering binding site similarity between different targets.