Why mosquito age matters?

Knowing the age of mosquito populations is important, as only old mosquitoes can transmit malaria. We developed an infrared spectroscopy/artificial intelligence approach that measures the age and species of wild mosquitoes in less than one minute. This technology can transform malaria surveillance.

Like Comment
Read the Article

How can we prevent malaria deaths?

Malaria kills 627,000 people every year, mostly children under 5 [1]. The main tool in the fight against malaria has been the control of the vector that transmits the disease: Anopheles mosquitoes. Vector control using insecticides has been very effective and averted 537 million clinical cases between 2000 and 2015 in Africa [2]. However, mosquito populations are becoming more and more resistant to these chemical-based interventions, which could make them less effective [3]. New hope comes from the methods that have been recently developed to attack mosquito populations [4]. However, before we can use these novel weapons, we still have a major challenge ahead.

What are the major challenges for the use of new malaria vector control?

While it is relatively straightforward to measure if newly developed vector control tools work in the lab, once they are deployed in the field it is very difficult to establish if they are working, and extensive time and resources are needed [5]. Indeed, malaria is a complex disease influenced by many factors, all difficult to quantify. However, if there were a quick and practical way to know if an intervention had reduced the mosquitoes that transmit malaria, this could drastically accelerate the implementation of new vector control strategies.

Why does mosquito age matter?

Knowing the distribution of ages of a mosquito population is very important, as only old female mosquitoes can transmit the deadly malaria parasites [6]. Therefore, to evaluate the success of a vector control intervention, we specifically need to know how it affected the older individuals. However, current methods of assessing mosquito age are slow and expensive as they are often based on mosquito dissections. In  this work we developed a solution that not only provides a rapid estimate of mosquito age but also of their species.

What is this new technology based on?

This new approach utilises infrared spectroscopy and artificial intelligence. By measuring the parts of the infrared light spectrum that a mosquito's surface absorbs (its cuticle), valuable information about its chemical composition can be obtained [7] (Figure 1).

Figure 1. A typical mosquito spectrum, showing absorbance at specific wavenumber/cm corresponding to chemical bonds present in the chitin, lipids or proteins of the cuticle.

As the cuticle of a mosquito changes during ageing and differs between species [8,9], we used this technique to rapidly obtain information on the chemical composition of the ageing cuticle. Then, we trained machine learning algorithms to identify the subtle differences between the spectra of young and old mosquitoes and of morphologically identical species. This artificial intelligence approach, which is similar to the one used for image recognition, allowed us to create models that can predict the age and species of mosquitoes based on their spectra, which can easily be obtained by placing mosquitoes in a spectrometer (Figure 2).

Figure 2. A scientist measuring a mosquito using a midinfrared spectrometer (left). The mosquito is placed in an anvil where the light passes through it and its absorbance is measured (right). The process takes less than a minute for each mosquito.

 How was this new technology evaluated?

We began by building a database of 40,000 mosquitoes of different known ages and species from the Anopheles gambiae complex reared in controlled laboratory conditions. We then trained machine learning algorithms to extract from those spectra the key signals associated with mosquito age and species. Finally, we refined those models by showing them wild mosquitoes collected in villages in Burkina Faso and Tanzania. To know if the age predicted by our algorithms was correct, we compared our predictions with age estimates of the same populations using the current gold-standard method, which is based on complex and time-consuming dissections of individual female mosquitoes and microscopic observations of the ovary morphology. These dissections can only be performed by highly trained technician and can take up to 30 minutes each, while our infrared-AI approach could be done by non-experts in less than a minute per mosquito. We found that both methods were very concordant, suggesting that this new technology works in wild mosquitoes (Figure 3).

Figure 3. Our approach that predicts mosquito age based on midinfrared spectroscopy and artificial intelligence (blue) was compared with the standard age-grading methods based on mosquito dissection and morphological assessment of the ovaries (yellow) in wild populations in Burkina Faso (a) and Tanzania (b). Figure from  Siria, Sanou, Mitton et al. Nature Communications 2022.

How will this new technology help malaria control?

By providing information on the age structure of mosquito populations, this approach would help malaria control by allowing to identify hotspots of potential malaria transmission and to determine if, and how well, vector control interventions are working. Our technology can be used as part of ongoing malaria surveillance efforts, where mosquitoes are routinely collected and stored in silica gel, similarly to our approach. The investment in this technology currently requires approximately $20,000 for the purchase of the spectrometer, and nothing else, apart for electricity for the instrument and a basic computer.

How was this technology developed?

The use of infrared spectroscopy to measure mosquito age was first published in 2009 by Valeliana S. Mayagaya, Floyd E. Dowell and other researchers at the Ifakara Health Institute [10]. However, those first attempts performed poorly on wild mosquito populations [11], possibly because they were based on the near-infrared region of the spectrum and on simple linear regression models. We modified this approach by looking at a different part of the spectrum, the mid-infrared region, where clearer biochemical signals could be detected and by applying innovative machine learning algorithms to learn from the complex patterns that constitute a spectrum. Importantly, our interdisciplinary team combined entomologists from malaria endemic countries in Tanzania at the Ifakara Health Institute (IHI) and in Burkina Faso at the Institut de Recherche en Sciences de la Santé (IRSS), who led the translation of our approach from the laboratory to the field. Instrumental to the development of this technology was the knowledge exchange between chemists, computing scientists and entomologists (Figure 4). 

Figure 4. One of the workshops on spectroscopy, machine learning and entomology  organised at the University of Glasgow, in partnership with the Institut de Recherche en Sciences de la Santé and the Ifakara Health Instititute. From left to right: Joshua Mitton, Francesco Baldini, Mario Gonzalez-Jimenez, Doreen Siria, Emmanuel Mwanga, Roger Sanou, Simon Babayan.

In the future, we aim to further improve this technology so that more precise predictions of mosquito age can made and extend this approach to other arthropods that carry human, animal and plant diseases.

References

  1. WHO. Word Malaria Report 2021. Word Malaria report Geneva: World Health Organization. (2021).
  2. Bhatt S, Weiss DJ, Cameron E, Bisanzio D, Mappin B, Dalrymple U, et al. The effect of malaria control on Plasmodium falciparum in Africa between 2000 and 2015. Nature 2015; 526 (7572): 207–11. doi: 10.1038/nature15535.
  3. Hemingway J, Ranson H, Magill A, Kolaczinski J, Fornadel C, Gimnig J, et al. Averting a malaria disaster: Will insecticide resistance derail malaria control? Lancet. 2016; 387 (10029): 1785–8. doi:10.1016/S0140-6736(15)00417-1.
  4. Barreaux P, Barreaux AMG, Sternberg ED, Suh E, Waite JL, Whitehead SA, et al. Priorities for Broadening the Malaria Vector Control Tool Kit. Trends Parasitol. 2017; 33 (10): 763–74. doi:10.1016/j.pt.2017.06.003
  5. Russell TL, Farlow R, Min M, Espino E, Mnzava A, Burkot TR. Capacity of National Malaria Control Programmes to implement vector surveillance: a global analysis. Malar J. 2020; 19 (1): 1–9. doi: 10.1186/s12936-020-03493-1
  6. Ohm JR, Baldini F, Barreaux P, Lefevre T, Lynch PA, Suh E, et al. Rethinking the extrinsic incubation period of malaria parasites. Parasit Vectors. 2018; 11 (178): 1–9. doi:10.1186/s13071-018-2761-4
  7. González Jiménez M, Babayan SA, Khazaeli P, Reedy E, Glew T, Doyle M, et al. Prediction of mosquito species and population age structure using mid-infrared spectroscopy and supervised machine learning. Wellcome Open Res. 2019; 4:76. doi:10.12688/wellcomeopenres.15201.3
  8. Caputo B, Dani FR, Horne GL, Petrarca V, Turillazzi S, Coluzzi M, et al. Identification and composition of cuticular hydrocarbons of the major Afrotropical malaria vector Anopheles gambiae s.s. (Diptera: Culicidae): Analysis of sexual dimorphism and age-related changes. J Mass Spectrom. 2005; 40 (12): 1595–604. doi:10.1002/jms.961.
  9. Suarez E, Nguyen HP, Ortiz IP, Lee KJ, Kim SB, Krzywinski J, et al. Matrix-assisted laser desorption/ionization-mass spectrometry of cuticular lipid profiles can differentiate sex, age, and mating status of Anopheles gambiae mosquitoes. Anal Chim Acta. 2011; 706 (1): 157–63. doi:10.1016/j.aca.2011.08.033.
  10. Mayagaya VS, Michel K, Benedict MQ, Killeen GF, Wirtz RA, Ferguson HM, et al. Non-destructive determination of age and species of Anopheles gambiae s.l. using near-infrared spectroscopy. Am J Trop Med Hyg. 2009; 81 (4): 622–30. doi:10.4269/ajtmh.2009.09-0192.
  11. Krajacich BJ, Meyers JI, Alout H, Dabiré RK, Dowell FE, Foy BD. Analysis of near infrared spectra for age-grading of wild populations of Anopheles gambiae. Parasit Vectors. 2017; 10 (1): 552. doi:10.1186/s13071-017-2501-1.

 

Authors

Francesco Baldini, Ph.D, holds a B.S. in biotechnology, a M.Sc. in medical biotechnology and a Ph.D in medical entomology. He is currently a Lecturer at University of Glasgow, Institute of Biodiversity Animal Health and Comparative Medicine. His current research interests focus on mosquito ecology and surveillance and malaria parasite-vector interactions.

Simon Babayan, Ph.D, holds a B.S. in zoology, ecology, and evolution, an M.Sc. in physiology and immunology, and PhD in parasitology and immunology. He is currently a Lecturer at University of Glasgow, Institute of Biodiversity Animal Health and Comparative Medicine. His current research interests focus on wild immunology and disease ecology.

Mario Gonzalez-Jimenez , Ph.D, holds a M.Sc. in Chemistry and a Ph.D in Physical Chemistry. He is currently a Research Associate at the School of Chemistry of the University of Glasgow. His current research interests focus on the use of vibrational spectroscopy to characterise the properties of biological matter.

 

 

 

Francesco Baldini

Lecturer, University of Glasgow

I am a vector biologist studying the ecology of Anopheles mosquitoes to fight malaria. My main research themes are:

  1. Determine mosquito life-history shifts in response to vector control.
  2. Understand Plasmodium falciparum-Anopheles interactions.
  3. Elucidate how endosymbionts influence human and avian malaria transmission.
  4. Develop new surveillance tools for disease vector populations.
  5. Sustain and optimize vector control strategies.
1 Contributions
0 Followers
0 Following