Scientists from the Jiangsu University of Science and Technology in Zhenjiang, China recently tested different analysis methods of using portable near-infrared (NIR) spectroscopy to detect Bombyx mori nuclear polyhedrosis virus within silkworms. Their findings were published in Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy (1).
silkworm cocoon on green mulberry leaves | Image Credit: © nd700 - stock.adobe.com
Raw silk production accounts for 82% of the world’s sericulture production. This process occurs when silkworms spin themselves into a cocoon, with the ultra-soft fibers being harvested from the cocoon in their raw state by being boiled in hot water and stirred until the cocoons unravel (2). However, this process has several problems to it, one of which being that silkworms can be easily infected by viruses. The most common and harmful virus that affects silkworms is Bombyx mori nuclear polyhedrosis virus (BmNPV). For early BmNPV diagnosis, conventional methods rely on human senses, which are highly sensitive to subjective influence that can lead to low accuracy. Alternate methods have been created for this purpose, such as the anti-viral antibody-based ELISA and nucleic acid probes; however, while highly sensitive, they can be high-cost, time-consuming, and labor-intensive. As such, fast, accurate, and effective early diagnosis methods are still needed.
For this study, a portable near-infrared (NIR) spectroscopy system was used to collect the spectra of normal silkworms and the infected silkworms induced by BmNPV administration. Unlike traditional NIR spectrometers, portable NIR spectrometers can perform on-site analysis without needing to transport samples back to a laboratory, which improves work efficiency and saves both time and money.
Various spectral methods were applied for the classification analysis; principal component analysis (PCA), linear discriminant analysis (LDA), and partial least squares discriminant analysis (PLSDA) were used. However, PCA and LDA were unable to achieve accurate classification results.
For the PLSDA calibration, after spectra pretreatment using a standard normal variate (SNV) transformation combined with second derivative, the technique yielded a highly accurate identification performance. The PLSDA approach resulted in low classification errors of 0.023, 0.033, and 0.030 for the calibration set, cross-validation set, and test set, respectively. To summarize these results, misclassification occurred 2.3% of the time for the calibration set; 3.3% of the time for the cross-validation set; and 3.0% of the time for the test set. All these factors were captured with higher sensitivity and specificity. These findings allowed scientists to identify BmNPV-infected silkworms using portable NIR spectroscopy; though there is more research to be done in this regard, this method has the potential to effectively reduce losses for the sericulture industry.
(1) Liu, Y.; Deng, S.; Li, Y.; Zhang, Y.; et al. Fast Identification of the BmNPV Infected Silkworms by Portable NIR Spectroscopy and Chemometrics. Spectrochim. Acta Part A: Mol. Biomol. Spectrosc. 2024, 314, 124158. DOI: https://doi.org/10.1016/j.saa.2024.124158
(2) Jackson, F. How Is Silk Made, And Is It Vegan? Allplants Ltd 2024. https://allplants.com/blog/lifestyle/how-is-silk-made-and-is-it-vegan (accessed 2024-4-8).
Machine Learning and Optical Spectroscopy Advance CNS Tumor Diagnostics
July 1st 2025A new review article highlights how researchers in Moscow are integrating machine learning with optical spectroscopy techniques to enhance real-time diagnosis and surgical precision in central nervous system tumor treatment.
Toward a Generalizable Model of Diffuse Reflectance in Particulate Systems
June 30th 2025This tutorial examines the modeling of diffuse reflectance (DR) in complex particulate samples, such as powders and granular solids. Traditional theoretical frameworks like empirical absorbance, Kubelka-Munk, radiative transfer theory (RTT), and the Hapke model are presented in standard and matrix notation where applicable. Their advantages and limitations are highlighted, particularly for heterogeneous particle size distributions and real-world variations in the optical properties of particulate samples. Hybrid and emerging computational strategies, including Monte Carlo methods, full-wave numerical solvers, and machine learning (ML) models, are evaluated for their potential to produce more generalizable prediction models.
Combining AI and NIR Spectroscopy to Predict Resistant Starch (RS) Content in Rice
June 24th 2025A new study published in the journal Food Chemistry by lead authors Qian Zhao and Jun Huang from Zhejiang University of Science and Technology unveil a new data-driven framework for predicting resistant starch content in rice