E-nose-based VOC analysis for rapid detection of grasserie disease in silkworms

The major threat associated with the silkworm industry is grasserie disease caused by the Bombyx mori nucleopolyhedrovirus (BmNPV), that leads to severe economic losses in sericulture. Current diagnostic technique are time consuming and requires skilled professionals, necessitating the development of a rapid, non-invasive detection system. This work details the use of an electronic nose (e-nose) device for detecting grasserie disease in silkworms based on volatile organic compound (VOC) analysis. The e-nose system consists of a sensor array with four sensors MQ-135, MQ-138, TGS-822, and TGS-2620 placed in a sensor chamber, with data acquisition performed using Arduino Uno Rev 3. Preprocessing was performed using Discrete Wavelet Transform (DWT), followed by Principal Component Analysis (PCA) for dimensionality reduction. K-Nearest Neighbors (KNN), Naïve Bayes, and Support Vector Machine (SVM) classifiers were employed for disease classification. The 10-fold cross-validation method with SVM yielded the best results, achieving an accuracy of 94.1 ± 0.5%, precision of 94.3 ± 0.4%, recall of 94.6 ± 0.5%, and F1-score of 94.4 ± 0.4%. This study demonstrates the potential of e-nose technology as a rapid, non-invasive, and cost-effective tool for detecting grasserie disease in silkworms. Further improvements in sensor stability and environmental robustness can enhance its real-world applicability for sericulture management.