Prediction of drug-target interactions via neural tangent kernel extraction feature matrix factorization model

Yu Wang, Yu Zhang, Jianchun Wang, Fang Xie, Dequan Zheng, Xiang Zou, Mian Guo, Yijie Ding, Jie Wan, Ke Han

Abstract: Highlights•Manually selected drug and target characteristics can lead to the loss of critical chemical information in drugs and proteins. Deep learning models can compensate for this deficiency, so this paper proposes a feature extraction logical neighborhood regularized matrix factorization method based on neural tangent kernel model. Taking advantage of automatic feature extraction by NTK model. This method uses neural tangent kernel model to train and matrices to obtain drug and target feature matrices. Experimental results show that the proposed method achieves good results on four benchmark datasets. This is verified to have good performance in predicting DTIs. In addition, this method is further applied to predict new DTIs and has achieved good results.