In 2022, Dr. Ian Kaniu a Senior Lecturer at University of Nairobi was awarded the Research and Innovation Grant by KENET in the Special Interest Group (SIG) for Computational Modelling and Materials Science (CMMS). The purpose of the grant was to lead the efforts in developing an on-site detection method for chlorothalonil in fresh vegetable /farm produce.  

The increasing use of pesticides in modern agriculture has raised serious concerns about food safety, environmental sustainability, and public health. Chlorothalonil, a widely used pesticide to control fungal diseases in crops, has been linked to health risks, including cancer. Dr. Kaniu's research aims to provide an innovative, cost-effective, and non-destructive solution for detecting pesticide residues using unique spectral fingerprints and machine learning models for rapid pesticide screening.  

Pesticide residues in food pose significant health risks, particularly in developing countries where regulatory enforcement may be weak. In Kenya, for example, recent studies have linked prolonged exposure to chlorothalonil with increased risks of cancer, endocrine disruption, and organ toxicity. Despite these concerns, conventional detection methods remain largely laboratory-based, time-consuming, and expensive. The lack of accessible screening tools has resulted in poor regulatory compliance and undetected exposure to harmful pesticide levels. 

A major challenge in pesticide residue detection is the need for fast and accurate analysis without compromising sensitivity and reliability. Traditional techniques such as gas chromatography (GC) and mass spectrometry (MS) offer high accuracy but require complex sample preparation and specialized laboratory infrastructure. These constraints highlight the necessity of portable, real-time detection solutions that can be used in farms, markets, and food processing facilities. 

Diffuse Reflectance Spectroscopy (DRS) is an optical technique that analyzes how light interacts with a sample. Unlike traditional spectroscopy methods that rely on transmission, DRS measures reflected light to extract unique spectral signatures associated with different chemical compounds. 

 

A Raman Spectrometer 

In the case of chlorothalonil detection, DRS enables the identification of unique spectral fingerprints corresponding to the pesticide's molecular structure. When light interacts with pesticide-contaminated produce, it produces a distinct spectral pattern that can be analyzed to determine the presence and concentration of pesticide residues. 

The advantages of DRS-based pesticide screening are truly game-changing. Unlike traditional methods, DRS offers non-destructive testing, meaning it doesn't require destroying samples or using extensive chemical preparations. This makes it not only faster but also more environmentally friendly. With results available in just minutes, DRS enables real-time monitoring, ensuring that pesticide residues can be detected quickly, even in the field. It's also incredibly cost-effective, as it minimizes the need for consumables, cutting operational costs. Perhaps most exciting of all, the portability of handheld DRS devices allows for on-site testing, making it a powerful tool for farmers, food safety inspectors, and regulatory bodies to monitor pesticide contamination wherever it’s needed. 

Portable spectrometer with integrated onboard processing for instant residue analysis.  

While DRS provides valuable spectral data, manual interpretation can be challenging and error-prone due to overlapping spectral signals from different substances. To overcome this limitation, Dr. Kaniu's team has integrated machine learning algorithms to enhance the accuracy and efficiency of pesticide residue identification. 

Machine learning models, particularly supervised learning techniques, are trained using datasets containing spectral information from pesticide-contaminated and pesticide-free samples. By analyzing patterns in the spectral data, the models can distinguish between contaminated and safe produce with high accuracy. 

Dr. Kaniu's team has achieved impressive success in training machine learning models to enhance pesticide detection. By leveraging advanced algorithms like Support Vector Machines (SVMs), they’ve developed a system that can effectively identify complex spectral patterns, ensuring precise results even in the most challenging samples. The team also uses Random Forest, a robust algorithm that can handle variability in spectral data, making it adaptable to real-world conditions. Additionally, Artificial Neural Networks (ANNs) are employed to uncover intricate relationships between spectral patterns and pesticide concentrations, delivering highly accurate and reliable results. These cutting-edge technologies are transforming pesticide screening, providing faster, more accurate solutions to ensure food safety and health. 

These models have demonstrated remarkable performance in classifying pesticide contamination with precision rates exceeding 95%. The combination of DRS and ML represents a breakthrough approach in achieving fast, reliable, and scalable pesticide screening solutions. 

Handheld NirQuest NIR spectrometer (in development) – refining detection algorithms through machine learning. 

Dr. Kaniu's research team is looking ahead with an ambitious vision to create a standalone detection system that will empower farmers, food safety inspectors, and policymakers to quickly and accurately monitor pesticide residues. Their future research focuses on several exciting developments: first, expanding their technology to detect a wider range of pesticides, ensuring a comprehensive approach to food safety. They also aim to miniaturize detection devices, making handheld and smartphone-integrated sensors more accessible for on-the-spot testing. Further enhancing this innovation, the team is exploring the integration of automated decision support systems, where cloud-based platforms will provide real-time alerts and actionable recommendations to stakeholders. Equally important is their commitment to regulatory collaboration, working closely with policymakers to incorporate DRS-ML-based detection methods into national food safety protocols. These steps represent a transformative leap toward a safer, more sustainable food system, one that benefits everyone from the field to the dinner table. 

The integration of Diffuse Reflectance Spectroscopy (DRS) and Machine Learning (ML) to detect pesticide residues is a breakthrough in food safety technology. This innovative approach offers a quick, accurate, and affordable way to screen for chlorothalonil residues, and it has the potential to completely change the way we monitor pesticides in agriculture. Looking ahead, the team is focused on scaling up this technology, improving detection models, and ensuring it becomes a standard tool used across food production and distribution networks. 

As concerns over pesticide residues continue to rise, the DRS-ML method is set to become an essential part of the solution. By making real-time, on-site pesticide detection a reality, this technology will not only help protect public health but also promote more sustainable farming practices. With continued research and investment, such as the KENET Grant, we’re getting closer to a future where food safety is better, faster, and more accessible, ensuring a safer and healthier food supply for everyone. 

Building on the strong collaborative foundation established through the KENET grant, it is important to highlight the tangible impact the project has had. Beyond the successful execution of the research activities, the grant played a pivotal role in advancing academic and professional growth among team members. For instance, Charles Ndung’u, who was a PhD student at the time of the project, has since completed his doctorate—a milestone that reflects the supportive research environment fostered by the grant. Additionally, the experience and outcomes of this initiative have positioned the team to pursue further funding opportunities, with efforts currently underway to secure additional grants to expand and deepen the project’s scope. This demonstrates the lasting influence of the initial support received from KENET.