基于人工智能的药物-靶标相互作用预测

Drug-target Interaction Prediction with Artificial Intelligence

关键词:  药物-靶标相互作用  深度学习  概率矩阵分解算法  靶标预测

Key words:  drug-target interaction  deep learning  probabilistic matrix factorization  target prediction

摘要:

目的 建立一个高效的药物-靶标相互作用预测分类模型，为生物实验提供有力的补充工具。方法 研究开发一种基于深度学习的方法来预测药物-靶标相互作用：通过引入高维分子指纹和蛋白质描述符，并应用概率矩阵分解算法生成负样本集，构建一个高效的药物-靶标相互作用预测分类模型。结果 与其他已报道的方法相比，本方法具有可比性或优越性，预测准确性、特异性、敏感性以及AUC值均>90%，提示该方法在药物靶标预测方面具有良好的应用前景。结论 人工智能深度学习模型以及概率矩阵分解算法的结合有助于解决药物-靶标相互作用预测精度低、负样本选择不合理等问题。

Abstract:

OBJECTIVE To build an efficient drug-target prediction classification model and provide a useful complementary tool for biological experiments. METHODS In this study, a deep learning-based method was developed to predict drug target interaction. By introducing high dimensional molecular fingerprints and protein descriptors, and subsequently applying a probability matrix decomposition algorithm to generate negative samples, a promising drug target interaction classification model was constructed. RESULTS The method was comparable or superior to previously developed methods against the test sets, achieving ˃90% accuracy, specificity, sensitivity, and AUC. This method represented a promising tool for drug target prediction. CONCLUSION The combination of artificial intelligence deep learning model and probabilistic matrix factorization algorithm can help to solve the problems of low prediction accuracy of drug-target interaction and unreasonable selection of negative samples.