Mosquito identification using infrared spectroscopy and chemometrics

Mosquito control interventions are more effective when informed by routine entomologic surveillance. Thus, accurate and rapid species identification remains a critical component of operational mosquito control. Current methods to identify adult mosquitoes rely chiefly on microscopic identification by trained personnel. In some larger mosquito control programs, molecular methods may be used for species or pathogen identification and advanced techniques (e.g., age-grading by ovarian dissection) may be used to further assess the mosquito population structure. Each of these methods are labor intensive and subject to a series of operator or laboratory errors. Therefore, there is a need for rapid and non-destructive species identification techniques that can be used on a scale that is ecologically, economically, and epidemiologically meaningful. Our current research aims to develop methods of biochemical discrimination between different mosquito species using infrared spectroscopy. Infrared spectroscopy is a sensitive, information rich technique that is capable of detecting a wide range of molecular signals ranging from subtle changes in protein secondary structure to transmembrane protein-lipid interactions. The resulting spectral data, when coupled with a numerical analysis (chemometrics) method, such as partial least squares discriminant analysis, may be used to classify mosquitoes by species or physiologic status. Herein, we have applied Fourier transform infrared (FT-IR) microspectroscopy to identify four container-inhabiting Aedes species (Ae. aegypti, Ae. albopictus, Ae. japonicus, and Ae. triseriatus) obtained from both field and laboratory conditions. At present, our FT-IR classification success rate using partial least squares discriminant analysis, when compared to identification by a trained entomologist, is 94.5-100%. This method, which is rapid and easy to use, has the potential to decrease the labor costs and time associated with mosquito species identification. Further development coupled with process automation may provide operationally useful methods for rapid identification of many mosquito species and their physiologic status