State Key Laboratory of Organic Geochemistry and Guangdong Key Laboratory of Environmental Resources Utilization and Protection, Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, Guangzhou 510640, China.
Solid Waste and Chemicals Management Center, Ministry of Ecology and Environment of the People's Republic of China, Beijing 100029, China.
Environ Sci Technol. 2024 Jan 30;58(4):1944-1953. doi: 10.1021/acs.est.3c08016. Epub 2024 Jan 19.
Tissue-to-blood partition coefficients () are key parameters for assessing toxicokinetics of xenobiotics in organisms, yet their experimental data were lacking. Experimental methods for measuring values are inefficient, underscoring the urgent need for prediction models. However, most existing models failed to fully exploit data from diverse sources, and their applicability domain (AD) was limited. The current study developed a multimodal model capable of processing and integrating textual (categorical features) and numerical information (molecular descriptors/fingerprints) to simultaneously predict values across various species, tissues, blood matrices, and measurement methods. Artificial neural network algorithms with embedding layers were used for the multimodal modeling. The corresponding unimodal models were developed for comparison. Results showed that the multimodal model outperformed unimodal models. To enhance the reliability of the model, a method considering categorical features, weighted molecular similarity density, and weighted inconsistency in molecular activities of structure-activity landscapes was used to characterize the AD. The model constrained by the AD exhibited better prediction accuracy for the validation set, with the determination coefficient, root mean-square error, and mean absolute error being 0.843, 0.276, and 0.213 log units, respectively. The multimodal model coupled with the AD characterization can serve as an efficient tool for internal exposure assessment of chemicals.
组织-血液分配系数（）是评估外来化合物在生物体内毒代动力学的关键参数，但缺乏其实验数据。测量值的实验方法效率低下，这突显了预测模型的迫切需求。然而，大多数现有模型未能充分利用来自不同来源的数据，其适用域（AD）有限。本研究开发了一种多模态模型，能够处理和整合文本（分类特征）和数值信息（分子描述符/指纹），以同时预测各种物种、组织、血液基质和测量方法的值。使用具有嵌入层的人工神经网络算法进行多模态建模。为了进行比较，还开发了相应的单模态模型。结果表明，多模态模型优于单模态模型。为了提高模型的可靠性，考虑了分类特征、加权分子相似性密度和结构活性景观中分子活性的加权不一致性的方法用于表征 AD。受 AD 约束的模型对验证集表现出更好的预测准确性，其确定系数、均方根误差和平均绝对误差分别为 0.843、0.276 和 0.213 个对数单位。结合 AD 特征的多模态模型可以作为化学品体内暴露评估的有效工具。
