Meta-analysis and machine learning to explore soil-water partitioning of common pharmaceuticals

Predicting sorption of diverse organic compounds in soil-water systems: Meta-analysis, machine learning modeling, and global soil mapping

In recent decades, the environmental detection of various organic compounds (OCs) has highlighted the limitations of conventional soil-water sorption models, which simplify complex experimental conditions and often overlook OCs with polyfunctional and ionizable structures. To address these shortcomings, we compiled a comprehensive soil-water sorption dataset encompassing 20,945 data points for 419 OCs with various functional groups and 1037 different soils. Meta-analysis of the dataset revealed the trends of soil sorption associated with OC substructures, soil properties, and solution conditions. Machine learning models employing the XGBoost algorithm, in conjunction with MACCS fingerprints and experimental conditions, were developed to cover the entire spectrum of speciation for cationic, neutral, and anionic species. Among these, the individual models tailored to each speciation achieved an overall root-mean-square-error value of 0.32 for log Kd. Model interpretation revealed that the models correctly understood the contributions of various substructures, such as multiple aromatic rings and nitrogen or oxygen atoms, to sorption. The models were also found to accurately capture isotherm nonlinearity and the pH effect on the sorption of ionizable OCs. Finally, utilizing soil properties from the Harmonized World Soil Database, the models predicted the sorption of diverse OCs based on global soil properties under simulated environmental scenarios.

A machine learning and experimental-based model for prediction of soil sorption capacity toward phenanthrene

Sorption by soil is the fundamental basis for environment fate of hydrophobic organic contaminants (HOCs), which varies significantly depending on diverse properties of soils. Therefore, a generalized approach to predict HOC sorption by soils is required. In this study, 488 data points were extracted from references and adopted to develop models for estimating the sorption capacities of phenanthrene in soils using six different machine learning (ML) approaches. The extreme gradient boosting (XGBT) model demonstrated the most favorable performance, achieving a coefficient of determination of 0.91 and root-mean-square errors of 0.24 for the testing dataset. The XGBT model's performance was further demonstrated by comparing with experimental data from batch sorption tests conducted on 20 soil samples collected from 17 provinces of China. The differences between the predicted values and the experimental values were statistically equal to zero (p = 0.14). Leveraging the XBGT model together with soil properties from the Harmonized World Soil Database, the distribution of sorption capacities in Chinese soils was successfully depicted on a national scale. This research is expected to contribute to a deeper understanding of the migration of persistent organic pollutants in terrestrial system. Furthermore, the established model holds implications for more precise and scientific soil environmental management.

A machine learning based approach for estimating site-specific partition coefficient Kd of organic compounds: Application to nonionic pesticides

The partitioning coefficient K_d for a specific compound and location is not only a key input parameter of fate and transport models, but also critical in estimating the safety environmental concentration threshold. In order to reduce the uncertainty caused by non-linear interactions among environmental factors, machine learning based models for predicting K_d were developed in this work based on literature datasets of nonionic pesticides including molecular descriptors, soil properties, and experimental settings. The equilibrium concentration (C_e) values were specifically included for the reason that a varied range of K_d corresponding to a given C_e occurred in a real environment. By transforming 466 isotherms reported in the literature, 2618 paired equilibrium concentrations of liquid-solid (C_e-Q_e) data points were obtained. Results of SHapley Additive exPlanations revealed that soil organic carbon, C_e, and cavity formation were the most important. The distance-based applicability domain analysis was conducted for the 27 most frequently used pesticides with 15952 pieces of soil information from the HWSD-China dataset by setting three C_e scenarios (i.e., 10, 100, and 1000 μg L^-1). It was revealed the groups of compounds showing log K_d < 0.06 and log K_d > 1.19 were composed mostly of those with log K_ow of -0.800 and 5.50, respectively. When log K_d varied between 0.100 and 1.00, it was impacted by interactions among soil types, molecular descriptors, and C_e comprehensively, which accounted for 55% of the total 2618 calculations. It could be concluded that site-specific models developed in this work are necessary and practicable for the environmental risk assessment and management of nonionic organic compounds.

Using Machine Learning to make nanomaterials sustainable

Sustainable development is a key challenge for contemporary human societies; failure to achieve sustainability could threaten human survival. In this review article, we illustrate how Machine Learning (ML) could support more sustainable development, covering the basics of data gathering through each step of the Environmental Risk Assessment (ERA). The literature provides several examples showing how ML can be employed in most steps of a typical ERA.A key observation is that there are currently no clear guidance for using such autonomous technologies in ERAs or which standards/checks are required. Steering thus seems to be the most important task for supporting the use of ML in the ERA of nano- and smart-materials. Resources should be devoted to developing a strategy for implementing ML in ERA with a strong emphasis on data foundations, methodologies, and the related sensitivities/uncertainties. We should recognise historical errors and biases (e.g., in data) to avoid embedding them during ML programming.