基于方差排序和深度神经网络的抗癌药物组合协同作用预测
Prediction of Anti-Cancer Drug Combinations Synergy Based on Variance Ranking and Deep Neural Network
张杉 1马敬山 2李玉双1
作者信息
- 1. 燕山大学理学院,河北秦皇岛 066004
- 2. 燕山大学理学院,河北秦皇岛 066004;石家庄邮电职业技术学院,河北 石家庄 050022
- 折叠
摘要
针对特定癌症类型,寻找协同的药物组合对于提高癌症疗效至关重要.基于方差排序和深度神经网络,构建了四类VarDNN模型,利用五折嵌套交叉验证和留一法,全面预测抗癌药物组合协同作用.实验结果显示:VarDNN模型的预测性能不仅优于已有经典模型,而且对于首次探讨的"新细胞系-新旧药物对"的预测结果也与已有结论相吻合.此外,VarDNN能够识别出与癌症发生发展密切相关的生物标志物,为抗癌药物组合筛选提供理论参考.
Abstract
Identifying novel synergistic combinations to a specific cancer is significant for improving cancer treatment.Based on variance ranking and deep neural network,four types of VarDNN models were built,and five-fold nested cross validation and leave-one-out method were employed to predict anti-cancer drug synergy comprehensively.The result shows that the prediction performance of VarDNN is superior to the state-of-the-art models.Most im-portantly,the synergy prediction of"new cell line-new and old drug pair",which has not yet been studied,is consistent with some known conclusions.Furthermore,VarDNN could identify some biomarkers closely related to the occurrence and development of cancers,and provide a theoretical reference for the screening of anti-cancer drug combinations in some extent.
关键词
抗癌药物组合/方差排序/深度神经网络/协同预测/生物标志物Key words
anti-cancer drug combination/variance ranking/deep neural network/synergy prediction/biomarker引用本文复制引用
基金项目
河北省自然科学基金(A2020203021)
河北省引进留学人员资助项目(C20200365)
出版年
2024
NETL
NSTL
万方数据