Abstract:Developing quantitative structure-activity relationships (QSARs) is an important approach to predicting the reactivity of HO radicals toward newly emerged organic compounds. As compared with molecular descriptors-based and the group contribution method-based QSARs, a combined molecular fingerprint-machine learning (ML) method can more quickly and accurately develop such models for a growing number of contaminants. However, it is yet unknown whether this method makes predictions by choosing meaningful structural features rather than spurious ones, which is vital for trusting the models. In this study, we developed QSAR models for the logkHO values of 1089 organic compounds in the aqueous phase by two ML algorithms—deep neural networks (DNN) and eXtreme Gradient Boosting (XGBoost), and interpreted the built models by the SHapley Additive exPlanations (SHAP) method. The results showed that for the contribution of a given structural feature to logkHO for different compounds, DNN and XGBoost treated it as a fixed and variable value, respectively. We then developed an ensemble model combining the DNN with XGBoost, which achieved satisfactory predictive performance for all three datasets: Training dataset: R-square (R2) 0.89–0.91, root-mean-squared-error (RMSE) 0.21–0.23, and mean absolute error (MAE) 0.15–0.17; Validation dataset: R2 0.63–0.78, RMSE 0.29–0.32, and MAE 0.21–0.25; and Test dataset: R2 0.60–0.71, RMSE 0.30–0.35, and MAE 0.23–0.25. The SHAP method was further used to unveil that this ensemble model made predictions on logkHO based on a correct 'understanding' of the impact of electron-withdrawing and -donating groups and of the reactive sites in the compounds that can be attacked by HO. This study offered some much-needed mechanistic insights into a ML-assisted environmental task, which are important for evaluating the trustworthiness of the ML-based models, further improving the models for specific applications, and leveraging the implicit knowledge the models carry.

A Quantitative Structure-Property Relationship Study on Reaction Rate Constants for Reductive Debromination of Polybrominated Diphenyl Ethers by Zero-Valent Iron

Quantitative structure-property relationship studies for direct photolysis rate constants and quantum yields of polybrominated diphenyl ethers in hexane and methanol.

[QSPR Studies on the Physicochemical Properties of Polybrominated Diphenyl Ethers Using Theoretical Descriptors Derived from Electrostatic Potentials on Molecular Surface].

A Theoretical Study on Reductive Debromination of Polybrominated Diphenyl Ethers

Quantitative Structure Activity Relationship Models For The Prediction Of Rat Lethal Dose 50% Of Aldehydes Based On Heuristic Method And Support Vector Machine

QSPR/QSAR models for prediction of the physicochemical properties and biological activity of polybrominated diphenyl ethers.

Excited States and Photodebromination of Selected Polybrominated Diphenyl Ethers: Computational and Quantitative Structure—Property Relationship Studies

Quantitative structure-property relationship study on physicochemical properties of polychlorinated diphenyl ethers and polybrominated diphenyl ethers

Reply to Comment by Sierra Rayne on “QSPR/QSAR Models for Prediction of the Physicochemical Properties and Biological Activity of Polybrominated Diphenyl Ethers” [chemosphere 66(10) (2007) 1998–2010]

Theoretical Studies On Structures, Properties And Dominant Debromination Pathways For Selected Polybrominated Diphenyl Ethers

Prediction of Supercooled Liquid Vapor Pressures and N-Octanol/air Partition Coefficients for Polybrominated Diphenyl Ethers by Means of Molecular Descriptors from DFT Method

Quantitative Structure–activity Relationship (QSAR) Models for Polycyclic Aromatic Hydrocarbons (pahs) Dissipation in Rhizosphere Based on Molecular Structure and Effect Size

Quantitative structure-activity relationship study on the degradation of polyhalogenated carbazoles by sulfidated zero-valent iron/peroxymonosulfate system

Molecular orbital studies on brominated diphenyl ethers. Part II--reactivity and quantitative structure-activity (property) relationships.

Predictive Insight into the Relationship Between AhR Binding Property and Toxicity of Polybrominated Diphenyl Ethers by PLS-derived QSAR

QSPR/QSAR models for prediction of the physico-chemical properties and biological activity of polychlorinated diphenyl ethers (PCDEs).

Development of quantitative structure gas chromatographic relative retention time models on seven stationary phases for 209 polybrominated diphenyl ether congeners

[Determination and QSPR Study on the Electrochemical Reduction Characteristics of Chlorinated Aromatic Compounds].

Quantitative Structure-Property Relationships for Octanol-Water Partition Coefficients of Polybrominated Diphenyl Ethers.

Shedding light on “Black Box” machine learning models for predicting the reactivity of HO radicals toward organic compounds

Qualitative and Quantitative Simulation of Androgen Receptor Antagonists: A Case Study of Polybrominated Diphenyl Ethers.