Predicting hydroxyl radical rate constants in water using a random forest model on a large dataset

IF 2.4 4区化学 Q3 CHEMISTRY, MULTIDISCIPLINARY

Journal of the Iranian Chemical Society Pub Date : 2026-10-12 DOI:10.1007/s13738-026-03659-9

Yali Xu, Yu Chen, Zhuo Wang, Saijin Huang, Xianwei Huang

{"title":"Predicting hydroxyl radical rate constants in water using a random forest model on a large dataset","authors":"Yali Xu, Yu Chen, Zhuo Wang, Saijin Huang, Xianwei Huang","doi":"10.1007/s13738-026-03659-9","DOIUrl":null,"url":null,"abstract":"<div><p>This study developed a high-accuracy random forest classification model to predict hydroxyl radical (·OH) reaction rate constants (log k<sub>OH</sub>) for a large-scale dataset of 1,159 organic compounds (the largest to date) in aqueous solution, achieving 87.4% test set accuracy and 97.5% overall accuracy. Beyond its predictive power, the model provides deep mechanistic insights into ·OH reactivity through 29 Dragon molecular descriptors. Analysis identified that electron-rich structural features, such as sulfur atoms (nS) and terminal sp² carbons (nR = Cp), enhance reactivity by facilitating radical addition and H-atom abstraction. Conversely, electron-withdrawing groups like halogens (nCL) and sulfonic acids (nSO2OH) suppress reactivity by reducing electron density at reaction sites. Steric accessibility (SpMin2_Bh(m)) and conjugated systems (JGI4) lower activation barriers by stabilizing transition states, while charge separation (JGI1) and electronegative atoms (O% and B01[O-O]) inhibit reactions. Unique halogen effects are observed: C-Br bonds (B02[C-Br]) sterically block ·OH access, whereas Cl-I bonds (B02[Cl-I]) unexpectedly enhance reactivity.</p><h3>Graphical abstract</h3><div><figure><div><div><picture><source><img></source></picture></div></div></figure></div></div>","PeriodicalId":676,"journal":{"name":"Journal of the Iranian Chemical Society","volume":"23 11","pages":""},"PeriodicalIF":2.4000,"publicationDate":"2026-10-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of the Iranian Chemical Society","FirstCategoryId":"92","ListUrlMain":"https://link.springer.com/article/10.1007/s13738-026-03659-9","RegionNum":4,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"CHEMISTRY, MULTIDISCIPLINARY","Score":null,"Total":0}

引用次数: 0

Abstract

This study developed a high-accuracy random forest classification model to predict hydroxyl radical (·OH) reaction rate constants (log k_OH) for a large-scale dataset of 1,159 organic compounds (the largest to date) in aqueous solution, achieving 87.4% test set accuracy and 97.5% overall accuracy. Beyond its predictive power, the model provides deep mechanistic insights into ·OH reactivity through 29 Dragon molecular descriptors. Analysis identified that electron-rich structural features, such as sulfur atoms (nS) and terminal sp² carbons (nR = Cp), enhance reactivity by facilitating radical addition and H-atom abstraction. Conversely, electron-withdrawing groups like halogens (nCL) and sulfonic acids (nSO2OH) suppress reactivity by reducing electron density at reaction sites. Steric accessibility (SpMin2_Bh(m)) and conjugated systems (JGI4) lower activation barriers by stabilizing transition states, while charge separation (JGI1) and electronegative atoms (O% and B01[O-O]) inhibit reactions. Unique halogen effects are observed: C-Br bonds (B02[C-Br]) sterically block ·OH access, whereas Cl-I bonds (B02[Cl-I]) unexpectedly enhance reactivity.

Graphical abstract

查看原文本刊更多论文

在大数据集上使用随机森林模型预测水中羟基自由基速率常数

本研究建立了一个高精度随机森林分类模型，用于预测水溶液中1,159种有机化合物（迄今为止最大的）的羟基自由基（·OH）反应速率常数（log kOH），测试集准确率达到87.4%，总体准确率达到97.5%。除了预测能力之外，该模型还通过29个Dragon分子描述符提供了对·OH反应性的深入机制见解。分析发现，硫原子（nS）和末端sp²碳（nR = Cp）等富电子结构特征通过促进自由基加成和h原子抽离来增强反应活性。相反，像卤素（nCL）和磺酸（nSO2OH）这样的吸电子基团通过降低反应部位的电子密度来抑制反应活性。空间可及性（SpMin2_Bh(m)）和共轭体系（JGI4）通过稳定过渡态降低了激活势垒，而电荷分离（JGI1）和电负性原子（O%和B01[O-O]）抑制了反应。观察到独特的卤素效应：C-Br键（B02[C-Br]）立体阻断·OH通路，而Cl-I键（B02[Cl-I]）意外地增强了反应性。图形抽象

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Journal of the Iranian Chemical Society 化学-化学综合

CiteScore

4.40

自引率

8.30%

发文量

230

审稿时长

5.6 months

期刊介绍： JICS is an international journal covering general fields of chemistry. JICS welcomes high quality original papers in English dealing with experimental, theoretical and applied research related to all branches of chemistry. These include the fields of analytical, inorganic, organic and physical chemistry as well as the chemical biology area. Review articles discussing specific areas of chemistry of current chemical or biological importance are also published. JICS ensures visibility of your research results to a worldwide audience in science. You are kindly invited to submit your manuscript to the Editor-in-Chief or Regional Editor. All contributions in the form of original papers or short communications will be peer reviewed and published free of charge after acceptance.