| Principal Investigator |
Neil F. Lobo, University of Notre Dame |
| International Collaborator(s) |
|
| Project Type |
Planning Grant |
| Project Title |
Validation of a scalable and automated mosquito species identification system |
| Priority Area |
Infectious diseases |
| The Intervention |
A vital part of our battle towards mosquito-borne disease elimination is the proper identification of the vector to species. Based on their behavior, each vector species may bite humans in distinct spaces and times, which impacts both where disease transmission happens, as well as how effective interventions are. Malaria programs utilize vector surveillance with parallel species identification – temporal and spatial evidence vital for decision making and intervention strategies. Morphological species identification is plagued by inadequate funding, human capacity (and training) as well as time required – with results often compromised by suboptimal sensitivity and specificity. There remains the need for scalable methods for quickly generating reliable, species-specific, high precision mosquito surveillance data. This study seeks to improve malaria vector identification using a visual algorithm-based approach. We propose to field-pilot and optimize a low-cost camera with an optimized computer vision system for malaria-vector identification. Partnering with IU in the Kenya highlands, this study seeks to utilize entomological samples in a tripartite analysis using morphological, molecular as well as the digital speciation platform – developed by Vectech, towards field testing, fine-tuning and validating the identification algorithm. Digital speciation data generated will be geospatially tagged, time-stamped, and up-loadable in near real-time to an analytics dashboard to guide malaria program planning and implementation. The optimization of this algorithm at the Kenyan site – with both characterized and novel species, will allows this scalable system to be applied in multiple geographies with diverse vectors – including vector surveillance systems across the USA. This is particularly important when considering the very real threat of introductions of novel disease vectors and the diseases they carry. The generation of a low-cost species identification tool, with both national and international applicability, will transform entomological surveillance present in research and government programs around the world. This would enable both an expansion of programs as well as the generation of better evidence in a cost-effective manner resulting in better decision-making. |
| Key Facilitators |
IUSM, VecTech |
| Target Population |
Human: Kenya highlands malaria endemic regions Mosquitoes: Vectors of malaria |
| Process to Implementation |
PiggyBac on a longitundian vector characterization effort. Mosquitoes will be identified by the system, fed into the machine learning system with aparallel molecular identifications towards validation and roundtruthing of algorithm. |
| Key Stakeholders |
Kenya MoH, ND, IUSM, VecTech, Indiana State Health Department |
| Scaled or Transferred? |
Scaled and transferred |
| Type of Research |
Longitudinal data collection with iteratative and adaptive sampling with parallel machine learning. |
| Published Materials |
N/A |
| Year Funded |
2021 |
