Many disease outbreaks start from viruses found in animals – but viruses in animals are very difficult to study. A Microsoft research project called Project Premonition aims to monitor these viruses using mosquitoes as natural field biologists that collect blood samples from animals.
It works like this: When mosquitoes bite animals, they obtain a small amount of blood containing genetic information about the animals that were bitten and viruses present. That information could be used to detect pathogens before they cause outbreaks.
Microsoft and its partners aim to utilize these insects’ work to identify where diseases come from and how they spread – and ultimately prevent outbreaks of new viruses – by using smart traps, drones and gene sequencing to capture mosquitoes and study the DNA they’ve collected.
To efficiently find and capture mosquitoes, the Project Premonition team has designed smart mosquito traps that use machine learning to selectively capture important species based on distinguishing wing-beat patterns. The traps also record environmental factors such as light, temperature and humidity – data that could be important in understanding how viruses are spreading. The team is applying drone technology to find mosquito hot spots and guide the placement of traps.
The system provides “a plethora of data we never had before about the behavior of the insects,” said Ethan Jackson, the Microsoft researcher who is leading Project Premonition.
The captured mosquitoes are turned into data by gene sequencing, producing more than 100 million pieces of small DNA sequences. These sequences are compared against the genomes of hundreds of thousands of organisms, from viruses and bacteria to reptiles and mammals.
This requires trillions of genetic comparisons, which can now be completed in just 12 hours using the computing power of the Microsoft Cloud. In the past, that data analysis took 30 days. The time savings of the Project Premonition system could help researchers detect new infectious diseases before they spread, enabling healthcare workers to get ahead of the curve in preparing responses.
The project benefits from recent developments in gene sequencing and computational biology that allow researchers to quickly search through samples for possible viruses, including specific diseases such as Zika, as well as ones that haven’t been discovered yet. It also helps researchers figure out which animals may be the sources and carriers of diseases.
“If we can detect these new viruses before they spread,” Jackson said, “we may someday prevent outbreaks before they begin.”