基于Fe-ZIF67纳米酶与机器学习的抗氧化能力分析研究
Analysis of Antioxidant Capacity Based on Fe-ZIF67 Nanozyme and Machine Learning
摘要
本研究采用表面配位修饰策略在ZIF67表面引入Fe2+，利用未配位咪唑氮原子锚定金属离子，制备了Fe-ZIF67纳米酶材料，通过TEM、XPS等手段表征了材料的形貌、结构和元素组成，XPS表征证实Fe2+成功锚定于ZIF67表面形成新的Fe-N活性中心。颜色变化值(B-B₀)/B₀与三种草本成分（MMP、LBP、DOPs）的浓度之间的相关系数均达0.99，证明该指标可作为反映抗氧化能力的量化指数用于后续分析。主成分分析结果显示，不同种类及浓度的样品在得分图中呈清晰聚类，验证了颜色信号数据的可区分性。将颜色图像和对应的抗氧化指数输入实验室已搭建的LSTM模型进行训练，测试集相关系数R2达0.95以上，实现了仅通过拍照即可对抗氧化能力进行准确预测。将该模型应用于真实露酒样品，所得抗氧化指数与样品中草本成分添加量趋势一致，验证了该方法在实际样品评估中的可行性。
Abstract
This study employed a surface coordination modification strategy to introduce Fe2+ onto the surface of ZIF67, using uncoordinated imidazole nitrogen atoms to anchor metal ions, thereby preparing Fe-ZIF67 nanozyme materials. The morphology,structure and elemental composition of the materials were characterised by TEM, XPS and other methods, with XPS confirming that Fe2+ was successfully anchored on the surface of ZIF67, forming new Fe-N active centres. The correlation coefficients between the colour change value (B-B)/B and the concentrations of three herbal components (MMP,LBP,DOPs) reached 0.99, demonstrating that this index can be used as a quantitative measure of antioxidant capacity for subsequent analysis. Principal component analysis showed that samples of different types and concentrations displayed clear clustering in the score plots, verifying the distinguishability of the colour signal data. By inputting colour images and the corresponding antioxidant indices into a laboratory-constructed LSTM model for training, the determination coefficient R2for the test set exceeded 0.95, achieving accurate prediction of antioxidant capacity using photographs alone. Applying this model to real tawny wine samples, the resulting antioxidant indices were consistent with the trends of herbal component additions in the samples, validating the feasibility of this method for evaluating actual samples.关键词
Fe-ZIF67/纳米酶/机器学习/LSTM/抗氧化能力/比色传感Key words
Fe-ZIF67/nanozyme/machine learning/LSTM/antioxidant capacity/colourimetric sensing引用本文复制引用
王子怡,单桂晔.基于Fe-ZIF67纳米酶与机器学习的抗氧化能力分析研究[EB/OL].(2026-05-21)[2026-05-24].https://chinaxiv.org/abs/202605.00187.学科分类
生物科学研究方法、生物科学研究技术/药学/计算技术、计算机技术
评论