分会

第三十四分会：环境化学

摘要

EDC-Predictor:一种整合药理学和毒性学谱预测内分泌干扰物的策略及其应用 俞卓杭，吴曾睿，李卫华，刘桂霞，唐赟* 华东理工大学药学院，上海市梅陇路130号，200237 *Email: ytang234@ecust.edu.cn 内分泌干扰物（EDC）被定义为在机体的生长、发育、生殖过程中，干扰体内血源性激素的正常生理过程而导致内分泌系统功能紊乱的外源性化合物。已有报道表明内分泌干扰不仅与这几个核受体（NR）相关，还可能与其它靶标蛋白相关，例如与激素合成相关的酶等。因此从系统药理学和计算毒理学两种角度来发现潜在的EDC可能是行之有效的策略。 在这个研究中，我们收集了1334个EDCs和1474个Non-EDCs。随后我们为这些化合物生成了四种类型特征：（1）基于化学结构信息计算分子指纹；（2）基于药物靶标相互作用构建基于网络的靶标谱；（3）基于大规模的毒理学数据构建基于机器学习的靶标谱；（4）整合两种靶标谱构建了组合靶标谱（CTP）。三种靶标谱的分析结果表明其能够区分NR相关的EDC、其他EDC、Non-EDC。基于这些分子特征使用了机器学习方法构建了计算预测模型，模型结果表明使用CTP构建的最优模型无论在五折交叉验证和外部验证上均表现最优。结合统计检验和通路富集分析，我们发现识别得到的关键靶标在内分泌干扰中发挥着重要作用。为了展示EDC-Predictor的实用价值，我们将其应用于两个案例研究，并将其与以前的预测工具进行了比较。实验结果表明EDC Predictor不仅识别更多机制的EDC，涵盖了NR和其他机制，同时在准确性上整体也表现更优。 综上所述，整合药理学谱和毒性学谱的新策略不仅预测NR相关的EDC，也能尝试从其他机制来发现EDC。同时本研究提出的策略也可以拓展到其他研究领域，如化合物的毒性效应预测等。 Fig. 1 Workflow of EDC Table1. The evaluation indicators of EDC prediction models using three features in test set validation. Feature Method Accuracy Precision Recall F1 score MCC AUC FP RF 0.824 0.833 0.779 0.806 0.646 0.900 NBTP XGB 0.851 0.865 0.806 0.835 0.700 0.909 CTP XGB 0.835 0.854 0.779 0.815 0.668 0.907 关键词：内分泌干扰物；网络推理；毒理学谱；药理学谱；计算预测； 参考文献 [1] Yu, Z.; Wu, Z.; M. Zhou.; Tang, Y. Environ. Sci. Technol. 2023, just accepted. [2] Wu, Z.; Lu, W.; Wu, D.; Tang, Y., Br. J. Pharmacol. 2016, 173, 3372. [3] Sun, L.; Yang, H.; Cai, Y.; Tang, Y. J. Chem. Inf. Model. 2019, 59, 973. EDC-Predictor: A novel strategy for prediction of endocrine-disrupting chemicals by integrating pharmacological and toxicological profiles Zhuohang Yu, Zengrui Wu, Weihua Li, Guixia Liu, Yun Tang* School of Pharmacy, East China University of Science and Technology, 130 Meilong Road, Shanghai, 200237 Endocrine-disrupting chemicals (EDCs) are defined as exogenous compounds that interfere with normal physiological processes of natural hormones during the growth, development and reproduction of the body, resulting in dysfunction of the endocrine system. Previous studies demonstrated that endocrine disruption is not only related to these nuclear receptors (NRs), but also related to other target proteins, such as enzymes related to hormone synthesis. Therefore, it may be an effective strategy to identify potential EDC from the perspectives of systems pharmacology and computational toxicology. In this study, we collected 1334 EDCs and 1474 Non-EDCs. Subsequently, we generated four features for these compounds: (1) calculating molecular fingerprints based on chemical structural information; (2) constructing a network-based target profiles based on drug-target interactions; (3) constructing machine learning-based target profiles based on large-scale toxicology data; (4) constructing combined target profiles (CTP) integrating two target profiles. The analysis results of three target profiles indicate that they can distinguish between NR-related EDCs, other EDCs, and Non-EDCs. Based on these features, prediction models were constructed using machine learning methods. The prediction results of models showed that the optimal model constructed by CTP performed best on the 5-fold cross validation and external validation. Combining statistical testing and pathway enrichment analysis, we found that identified key targets play an important role in endocrine disruption. To demonstrate the practical value of EDC-Predictor, we applied it in two case studies and compared it with previous prediction tools. The results indicate that the EDC-Predictor not only assist in identification of EDCs with more types of mechanisms, but also performs better in accuracy. In summary, the novel strategy of integrating pharmacological and toxicological profiles not only predicts NR-related EDCs, but also identifies EDCs from other mechanisms. Meanwhile, this strategy can be easily extended to other research fields, such as toxic effect prediction of compounds.

关键词

内分泌干扰;网络推理;毒理学谱;药理学谱;计算预测

线上墙报仅限年会已缴费参会代表观看。

您还没有登录，请您先 点击这里登录

陈龙 华东理工大学 共1人点赞了这篇线上墙报。