Effects of sample preservation methods and duration of storage on the performance of mid-infrared spectroscopy for predicting the age of malaria vectors

Résumé

Abstract Background Monitoring the biological attributes of mosquitoes is critical for understanding pathogen transmission and estimating the impacts of vector control interventions. Infrared spectroscopy and machine learning techniques are increasingly being tested for this purpose, and can accurately predict the age, species, blood-meal sources, and pathogen infections in Anopheles and Aedes mosquitoes. Since the techniques are still in early-stage development, there are no standardized procedures for handling mosquito samples. We therefore assessed the effects of different preservation methods and storage durations on the performance of mid-infrared spectroscopy for age-grading malaria-transmitting mosquitoes. Methods Laboratory-reared Anopheles arabiensis (N = 3,681) were collected as 5 or 17-day olds and killed with ethanol then preserved using either silica desiccant at 5°C, freezing at -20°C, or absolute ethanol at room temperatures. For each preservation method, the mosquitoes were divided into three groups and stored for 1, 4 or 8 weeks, then scanned using attenuated total reflection-Fourier transform infrared spectrometer on the mid-infrared wavelengths. Supervised machine learning classifiers were trained with the infrared spectra, and used to predict the mosquito ages. Results The classification of mosquito ages (as 5 or 17-day olds) was most accurate when the samples used to train the models (training samples) and samples being tested (test samples) were preserved the same way or stored for equal durations. However, when the test and training samples were handled differently, the classification accuracies declined significantly. Support vector machine classifiers (SVMs) trained using spectra of silica-preserved mosquitoes achieved 95% accuracy when predicting the ages of other silica-preserved mosquitoes, but this declined to 72% and 66% when age-classifying mosquitoes preserved using ethanol and freezing respectively. Similarly, models trained on one-week stored samples had declining accuracies of 97%, 83% and 72% when predicting ages of mosquitoes stored for 1, 4 or 8 weeks respectively. Conclusions When using mid-infrared spectroscopy and supervised machine learning to age-grade mosquitoes, the highest accuracies are achieved when the training and test samples are preserved the same way and stored for the same durations. Protocols for infrared-based entomological studies should therefore emphasize standardized sample-handling procedures; and possibly, additional statistical procedures such as transfer learning for greater accuracy.

Citer ce document

Accès au document

Ce document est hébergé sur son dépôt institutionnel d'origine.

Auteur(s)

Statistiques

Consultations : 1

Téléchargements : 0