Predictive modeling of estrogen receptor agonism, antagonism, and binding activities using machine- and deep-learning approaches.
Predictive modeling of estrogen receptor agonism, antagonism, and binding activities using machine- and deep-learning approaches.
复制标题
DOI：
10.1038/s41374-020-00477-2
复制
发表时间：
2021-04
期刊：
影响因子：
--
通讯作者：
Zhu H
中科院分区：
文献类型：
--
作者：
Ciallella HL;Russo DP;Aleksunes LM;Grimm FA;Zhu H
根据世界卫生组织的定义，内分泌干扰物是一种外源性物质或混合物，它改变内分泌系统的功能(S)，从而对完整的有机体、其后代或(亚)种群造成不利的健康影响。识别引起内分泌紊乱的毒物的传统实验检测方案可能既昂贵又耗时。计算机模拟已经成为筛选和优先选择潜在内分泌活性化合物的一种有前景和成本效益的替代方法。有效地识别合适的化学描述符和机器学习算法，包括深度学习，对于计算毒理学研究来说是一个相当大的挑战。在这里，我们试图将经典的机器学习算法和深度学习方法应用于一个由7,500多种化合物组成的小组，对与核雌激素受体(ERα和ERβ)活性相关的18种毒性预测方法(ToxCast)进行测试。本研究使用三种二值指纹(扩展连通性指纹、功能连通性指纹和分子接入系统)作为化学描述符。每个描述符与四个机器学习和两个深度学习(普通和多任务神经网络)方法相结合，以构建所有18个ER分析的模型。使用从五次交叉验证过程中获得的接收工作曲线(AUC)值下的面积来评估所得到的模型性能。结果表明，各个模型的AUC值在0.56到0.86之间。使用另外两组化合物(n=592和n=966)进行了外部验证，通过实验证明了它们与核ER的相互作用。每种化合物的激动剂、拮抗剂或结合分数是通过在相关分析模型中对其预测概率进行平均来确定的，作为外部验证，产生的AUC值范围从0.63到0.91。结果表明，多任务神经网络在对机械相关的端点进行建模时具有优势。基于单个模型平均值的共识预测仍然是计算毒性评估的最佳建模策略。
As defined by the World Health Organization, an endocrine disruptor is an exogenous substance or mixture that alters function(s) of the endocrine system and consequently causes adverse health effects in an intact organism, its progeny, or (sub)populations. Traditional experimental testing regimens to identify toxicants that induce endocrine disruption can be expensive and time-consuming. Computational modeling has emerged as a promising and cost-effective alternative method for screening and prioritizing potentially endocrine active compounds. The efficient identification of suitable chemical descriptors and machine learning algorithms, including deep learning, is a considerable challenge for computational toxicology studies. Here, we sought to apply classic machine learning algorithms and deep learning approaches to a panel of over 7,500 compounds tested against 18 Toxicity Forecaster (ToxCast) assays related to nuclear estrogen receptor (ERα and ERβ) activity. Three binary fingerprints (Extended Connectivity FingerPrints, Functional Connectivity FingerPrints, and Molecular ACCess System) were used as chemical descriptors in this study. Each descriptor was combined with four machine learning, and two deep learning (normal and multitask neural networks) approaches to construct models for all 18 ER assays. The resulting model performance was evaluated using the area under the receiving operating curve (AUC) values obtained from a five-fold cross-validation procedure. The results showed that individual models have AUC values that range from 0.56 to 0.86. External validation was conducted using two additional sets of compounds (n=592 and n=966) with established interactions with nuclear ER demonstrated through experimentation. An agonist, antagonist, or binding score was determined for each compound by averaging its predicted probabilities in relevant assay models as an external validation, yielding AUC values ranging from 0.63 to 0.91. The results suggest that multitask neural networks offer advantages when modeling mechanistically-related endpoints. Consensus predictions based on the average values of individual models remain the best modeling strategy for computational toxicity evaluations.
登录
查看更多内容
影响因子：
7.4
作者：
Attene-Ramos, Matias S.;Miller, Nicole;Huang, Ruili;Michael, Sam;Itkin, Misha;Kavlock, Robert J.;Austin, Christopher P.;Shinn, Paul;Simeonov, Anton;Tice, Raymond R.;Xia, Menghang
通讯作者：
Xia, Menghang
影响因子：
2.9
作者：
HAJI, M;KATO, KI;IBAYASHI, H
通讯作者：
IBAYASHI, H
通过使用各种化学信息学方法，对药物的人类口服生物利用度的批判性评估。
影响因子：
3.7
作者：
Kim, Marlene T.;Sedykh, Alexander;Chakravarti, Suman K.;Saiakhov, Roustem D.;Zhu, Hao
通讯作者：
Zhu, Hao
影响因子：
11.4
作者：
Browne, Patience;Judson, Richard S.;Thomas, Russell S.
通讯作者：
Thomas, Russell S.
影响因子：
4.1
作者：
Ciallella, Heather L.;Zhu, Hao
通讯作者：
Zhu, Hao