Application and Progress of Machine Learning in Pesticide Hazard and Risk Assessment

Long-term exposure to pesticides is associated with the incidence of cancer. With the exponential increase in the number of new pesticides being synthesized, it becomes more and more important to evaluate the toxicity of pesticides by means of simulated calculations. Based on existing data, machine learning methods can train and model the predictions of the effects of novel pesticides, which have limited available data. Combined with other technologies, this can aid the synthesis of new pesticides with specific active structures, detect pesticide residues, and identify their tolerable exposure levels. This article mainly discusses support vector machines, linear discriminant analysis, decision trees, partial least squares, and algorithms based on feedforward neural networks in machine learning. It is envisaged that this article will provide scientists and users with a better understanding of machine learning and its application prospects in pesticide toxicity assessment.

A deep learning-enabled smartphone platform for rapid and sensitive colorimetric detection of dimethoate pesticide

A novel deep learning-enabled smartphone platform is developed to assist a colorimetric aptamer biosensor for fast and highly sensitive detection of dimethoate. The colorimetric determination of dimethoate is based on the specific binding of dimethoate and aptamer, which leads to the aggregation of AuNPs in high-concentration NaCl solution, resulting in an obvious color change from red to blue. This color change provides sufficient data for self-learning enabled by a convolutional neural network (CNN) model, which is established to predict dimethoate concentration based on images acquired from a smartphone. To enhance user-friendliness for non-experts, the CNN model is then embedded into a smartphone app, enabling offline detection of dimethoate pesticide in real environments within just 15 min using a pre-configured colorimetric probe. The developed platform exhibits superior performance, achieving a regression coefficient of 0.9992 in the concentration range of 0-10 μM. Moreover, the app's performance is found to be consistent with the ELISA kit. These remarkable findings demonstrate the potential of combining colorimetric biosensors with smartphone-based deep learning methods for the development of portable and affordable tools for pesticide detection.

Detection of Pesticide Residues in Vegetables Sold in Changchun City, China

ABSTRACT

We evaluated fresh vegetables for residues of 18 pesticides with different chemical structures, including organochlorine pesticides, organophosphorus pesticides, carbamate pesticides, and pyrethroid pesticides and estimated that the potential health risks for consumers. A total of 313 samples were collected from 12 kinds of vegetables in Changchun, the capital of Jilin Province, People's Republic of China. Pesticide residues were analyzed by gas chromatography and mass spectrometry, and the curves were highly linear at 0.01 to 1.00 μg/mL (R2 ≥ 0.99). The mean recovery rate of the pesticides was 62 to 110% (relative standard deviation of <5%). The limit of detection was 0.0001 to 0.0167 mg/kg, the limit of quantification was 0.0002 to 0.0556 mg/kg, and the overall detection rate was 28.43%. The prevalence of pesticides and of samples above the standard limit were highest in celery, the prevalence of pesticides was lowest in potatoes, and the prevalence of samples above the standard limit was lowest in cucumber. Three of the 18 pesticides were not detected: omethoate, chlorpyrifos, and fenvalerate. Among the 15 pesticides detected, the maximum risk factor of six (carbofuran, omethoate, phorate, dicofol, dimethoate, and dichlorvos) is >1, indicating possible harm to human health. Residues of a single pesticide may not adversely affect a person's health, but multiple pesticide residues could present a health risk.

HIGHLIGHTS

Fast and Low-Cost Surface-Enhanced Raman Scattering (SERS) Method for On-Site Detection of Flumetsulam in Wheat

The pesticide residues in agri-foods are threatening people’s health. This study aims to establish a fast and low-cost surface-enhanced Raman scattering (SERS) method for the on-site detection of flumetsulam in wheat. The two-step modified concentrated gold nanoparticles (AuNPs) acted as the SERS substrate with the aid of NaCl and MgSO₄. NaCl is served as the activator to modify AuNPs, while MgSO₄ is served as the aggregating agent to form high-density hot spots. The activation and aggregation are two essential collaborative procedures to generate remarkable SERS enhancement and achieve the trace-level detection of flumetsulam. This method exhibits good enhancement effect with an enhancement factor of 10⁶ and wide linear range (5–1000 μg/L). With simple pretreatment, the flumetsulam residue in real wheat samples can be successfully detected with the limit of detection (LOD) down to 0.01 μg/g, which is below the maximum residue limit of flumetsulam in wheat (0.05 μg/g) set in China. The recovery of flumetsulam residue in wheat ranges from 88.3% to 95.6%. These results demonstrate that the proposed SERS method is a powerful technique for the detection of flumetsulam in wheat, which implies the great application potential in the rapid detection of other pesticide residues in various agri-foods.

Residue Analysis of Insecticides in Potatoes by QuEChERS-dSPE/UHPLC-PDA

Insecticides are broadly applied in agriculture to defend crops from illnesses and pest attacks, consequently guaranteeing high production. However, their residual deposits in food products are becoming a main concern with regard to human consumption. As such, sensitive analytical methods should be developed to assess, prevent and control insecticide residues. In this research, an accurate, fast and reliable residual analytical method, that is quick, easy, cheap, effective, rugged and safe, combined with dispersive solid phase extraction (QuEChERS-dSPE), was developed for the determination of the most common insecticides used in potatoes cultivation (chlorpyrifos, λ-cyhalothrin, deltamethrin and acrinathrin), using an ultra-high performance chromatography photodiode array detector (UHPLC-PDA). The most influential extraction and instrumentation parameters that affect the method’s performance, such as extraction solvent, ratio salts, sorbents, stationary phases, gradient conditions and eluents, were assessed. Under the ideal conditions, good linearity (0.992–0.998), limits of detection (0.02–0.47 µg/kg) and quantification (0.06–1.58 µg/kg), recovery (94.1 to 112%) and precision (relative standard deviation <18%) were achieved for spiked levels between 2.5 and 50 µg/kg. The obtained results revealed that the potatoes analyzed do not represent any concern for human healthy, as the insecticide residues detected were lower than the maximum residue limits set by the European Union, Codex Alimentarius, and other organizations.

Monitoring of the Organophosphate Pesticide Chlorpyrifos in Vegetable Samples from Local Markets in Northern Thailand by Developed Immunoassay

Chlorpyrifos is an organophosphate pesticide that is wildly used among farmers for crop protection. However, there are concerns regarding its contamination in the environment and food chain. In the present study, an in-house indirect competitive enzyme-linked immunosorbent assay (ic-ELISA) specific for detecting chlorpyrifos is developed and validated against gas chromatography–flame photometric detection (GC-FPD) as the conventional method. The developed ic-ELISA was used for detecting chlorpyrifos residue in vegetable samples. The developed ic-ELISA showed good sensitivity to chlorpyrifos at an IC₅₀ of 0.80 µg/kg, with low cross-reactivity to other organophosphate pesticides. The 160 samples were collected from local markets located in the Chiang Rai, Chiang Mai, and Nan provinces in northern Thailand. The positive rate of chlorpyrifos residues in the vegetable samples was 33.8%, with the highest levels found in cucumbers, coriander, and morning glory, at 275, 145, and 35.3 µg/kg, respectively. The highest median levels of chlorpyrifos found in the detected samples were Chinese cabbage (332 μg/kg), cucumber (146.3 μg/kg) and Chinese Kale (26.95 μg/kg). The developed ic-ELISA is suitable for the rapid quantitation of chlorpyrifos residues.

Evaluation of Pesticide Residue Dynamics in Lettuce, Onion, Leek, Carrot and Parsley

The dynamics of 32 active substances contained in pesticide formulations (15 fungicides and 17 insecticides) were analyzed in iceberg lettuce, onion, leek, carrot, and parsley. Pesticide residues were monitored from the time of application until harvest. In total, 114 mathematical models of residue dissipation were developed using a first-order kinetic equation. Based on these models, it was possible to predict the action pre-harvest interval (the time between the last pesticide application and crop harvest) needed to attain a targeted action threshold (value significantly lower than the maximum limit) for low-residue vegetable production. In addition, it was possible to determine an action pre-harvest interval based on an action threshold of 0.01 mg kg-1 to produce vegetables intended for zero-residue production. The highest amount of pesticide residues were found in carrot and parsley leaves several days after treatment, and pesticide dissipation was generally slow. Lower amounts were found in leeks and lettuce, but pesticide dissipation was faster in lettuce. According to our findings, it seems feasible to apply reduced pesticide amounts to stay below unwanted residue levels. However, understanding the effectivity of reduced pesticide application for controlling relevant pest organisms requires further research.

Rapid detection of trichlorfon residues by a microfluidic paper-based phosphorus-detection chip (μPPC)

We propose a novel microfluidic paper-based phosphorus-detection chip (μPPC) for rapid analysis of trichlorfon, one of the most common organophosphate pesticide residues.