Machine learning-based analysis on retention time of pesticides for eco-friendly and cost-effective analytical method development
Year: 2025
Type: article
Abstract: Developing eco-friendly and cost-effective analytical methods is one of the essential requirements in the present scenario. A study on the role of physicochemical properties/structural features of the chemicals/molecules for determining retention time/mobility will significantly impact the solvents/mobile phase selection for the analysis. Currently, various techniques, including computational and machine learning, are employed to study the correlations between the structural features and the analytical results of chemicals. In the present investigation, a machine learning-based classification analysis was carried out to understand the importance of structural features of pesticides (324 pesticides) for their retention time using support vector machine (SVM) and Naïve Bayes (NB) algorithms. The physicochemical descriptors and MACCS fingerprints of the pesticides were calculated using PaDEL software, and their analysis was performed using WEKA software. The reported retention time of >21 min was considered a high retention time (poor eluting power), and <21 was considered a low retention time (good eluting power). The results derived from the analysis showed that the models developed with a complete training set (all compounds) and validated by test set (30%) and 5-fold cross-validation methods provided significant statistical parameters such as accuracy, precision, specificity, Mathew’s correlation coefficients (MCC), enrichment factor, F-score, etc. The contributed physicochemical descriptors/MACCS fingerprints revealed that the presence of one six-membered/aromatic ring and =O containing a ring in the pesticides have low retention time. In comparison, more than one aromatic ring increases the retention time. The presence of double bonds, sulfoxide groups, heterocycles with oxygen atoms, sulfur connected with heteroatoms, ring systems, and more heteroatoms other than nitrogen increases the retention time of the molecules. Further, the presence of the amino group, nitrogen atom present in chain/heterocycle, phosphorus, sulfur, tertiary groups, more than one methyl/methoxy group, etc, are responsible for reducing the retention time of the molecules. These results revealed that selecting solvents with polar properties may elute the pesticides quickly, reducing the retention time. The decreased retention times of the pesticides may improve the eco-friendly analysis and reduce the analysis cost, subsequently reducing the price of the pesticides.
Authors: N.S. Hari Narayana Moorthy, Shivani Singh Baghel, Nisha Porte, Sarbas Hakkeem, Akhilesh Tiwari, C. Karthikeyan
Institution: Indira Gandhi National Tribal University
Language: English
Subfield: Computational Theory and Mathematics
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $2,000
OpenAlex ID: W4413190248