A Microsoft AI for Earth grant will help researchers bust the spread of infection for millions of people sustainably
A car trundles along a bumpy village road in the Fiji Islands in the South Pacific. As it passes through the lush tropical landscape, a woman in the backseat reaches into a box full of sealed plastic tubes. She opens them one-by-one and methodically shakes the contents out through an open window.
She is releasing a batch of laboratory-bred mosquitoes that have been modified by scientists to eliminate the scourge of dengue fever.
Regular releases like this by field workers and volunteers are producing dramatic trial results in selected communities in Asia Pacific and South America. Now researchers are turning to digital technologies to scale up their fight worldwide.
Around 40 percent of the world’s population – about 3 billion people in 100 countries – live in communities with a risk of dengue and other potentially deadly mosquito-borne viruses like Zika, yellow fever, and chikungunya.
Many of those people struggle with poverty and overcrowding, and tragically the most vulnerable to these diseases are often the very young.
“The haunting memory of dengue is the body ache,” recalls Evisake Wainiqolo, a Fijian mother of seven, who was infected as a child. “The pain is beyond explanation.”
There are no cures. But what if something could curb the power of mosquitoes to infect people?
Well, there is. It’s a bacteria called Wolbachia. And, in a way, it is to mosquitoes what kryptonite is to Superman. That’s because Wolbachia limits the replication of dengue and those other viruses within a mosquito’s body.
Using techniques developed by scientists at Monash University in Australia, the World Mosquito Program breeds mosquitoes with Wolbachia-infused cells and releases them into the environment to mate with local mosquitoes. This interbreeding spreads Wolbachia across entire mosquito populations and neutralizes their disease-carrying capabilities.
The common house mosquito, Aedes aegypti
Years of research, some trial-and-error deployment, and lots of labor-intensive grassroots campaigning have delivered impressive results in targeted communities.
Now the Program – a global not-for-profit research consortium headquartered in Vietnam – is planning to dramatically scale up its ambitions and scope. It has just been awarded a Microsoft AI for Earth grant to make Wolbachia’s disease-busting impact global.
With the help of data, machine learning, artificial intelligence (AI), and the computation power of the cloud, this humble micro-organism could soon become an international public health superhero.
The key to success is determining the best release points for modified mosquitoes to maximize impact, says Ben Green, Senior Project Delivery Manager at the Program, which has been working toward a goal of protecting 100 million people in 12 countries.
Its researchers are now pulling together massively detailed data sets to create a predictive deep learning model that will determine the best release points anywhere in the world.
AI does its best when it just kind of disappears into the background and lets people get on with the task at hand – Lucas Joppa, Chief Environmental Officer at Microsoft
Lucas Joppa, Chief Environmental Officer at Microsoft, says AI can change the game for not-for-profits. In this case, the Program’s AI model will have the potential to turbo-charge its work and impact by freeing up its researchers from the painstaking and time-consuming task of analyzing data.
“They will no longer be governed by the speed and scale of how they do their data analytics. AI does its best when it just kind of disappears into the background and lets people get on with the task at hand,” Joppa says. “They can pull together all of the data, all the expertise available, and expand their operations across the world instead of going from project to project, location to location.”
How Wolbachia works
Wolbachia is a naturally occurring type of bacteria that lives inside the cells of around 60 percent of insect species—but crucially not in Aedes aegypti, the mosquito whose bites can infect people.
But when scientists introduce Wolbachia into the cells of an Aedes aegypti, the mosquito’s virus-spreading capabilities are dramatically reduced. On top of that, Wolbachia also messes up a mosquito’s love life.
When a male with Wolbachia mates with a female that does not have it, the eggs she lays will not hatch. When a female with Wolbachia mates with a male who does not have it, her eggs will all produce offspring with Wolbachia. When both a male and a female have Wolbachia, then again, all their offspring will have it.
Within a few generations, the number of mosquitoes with Wolbachia exponentially multiplies until nearly all have it. The result: people will still have to put up with mosquito bites, but their communities will become free of those diseases.
In recent years, researchers at the Program have perfected ways of introducing Wolbachia into mosquito cells and are now busy breeding modified mosquitoes that are being released in a number of communities around the world.
A boy in Puducherry in southern India unscrews a jar and shakes out dozens of mosquitoes as he walks down a street in his impoverished neighborhood. In a village near Yogyakarta, on the Indonesian island of Java, a woman pours water carrying mosquito eggs into a pond.
The same sorts of scenes happen in targeted communities in Colombia and Brazil. And in Vietnam, a man pops the lid off a plastic tube as he sits on the back of a motor scooter in the market town of Vinh Luong.
“When we release these mosquitoes, it’s like we are releasing hope,” explains Samu Tuidraki, the chief of Narewa, a village in Fiji.
We will be able to release Wolbachia mosquitoes where they will have the most effect with analysis at a countrywide scale instead of at a neighborhood scale. – Ben Green, Senior Project Delivery Manager, World Mosquito Program
Australia’s Far North Queensland has been declared essentially dengue-free for the first time in about a century, following an intensive release campaign. Other targeted efforts are making strong progress in Asia and South America, where authorities have long tried to wipe out mosquito populations with insecticides.
“Our Wolbachia method is natural and self-sustaining,” Green says. “As a large-scale public health intervention, we believe that this a cost-effective way. The evidence so far is that it can sustain itself in local populations for up to seven years. And we expect it will continue.”
How machine learning and AI will help take the fight global
The Program’s data science partner, Gramener, is developing machine learning for the AI model. It will tap the Program’s existing release point records as well as many other datasets on human population densities, land use, industrial sites, weather, and other variables. Satellite imagery will be a big part of mapping out large urban areas with strategic and granular accuracy.
The aim is to have the ability to pinpoint multiple impactful release points within blocks of as little as 100 square meters.
“We want to target the areas where our intervention is needed most,” Green says. “We will be able to release Wolbachia mosquitoes where they will have the most effect with analysis at a countrywide scale instead of at a neighborhood scale. Our ambition is to be able to look at a whole country and run the model over all its urban areas and let it give an unprecedented snapshot of where we can have the most impact.”
Joppa says machine learning and AI are potent tools for not-for-profits that want to tackle big challenges but have limited resources.
“The World Mosquito Program started with the objective of figuring out how to attack a problem. In this case, they worked out how to neutralize the disease-carrying ability of mosquitoes. Then they worked out where they needed to release these mosquitoes.”
“They started collecting tons of data. It then became a really messy data problem as they tried to compare a bunch of different data sets to work out where they could be most efficient.”
“Ultimately, this is where machine learning comes in. It allows you to take all of that data, abstract it down to a single estimate of probability and map it out. It is cost-effective, and it is super scalable. Instead of figuring out data visualization and analytics for one particular area, you can now do it for an entire city, for an entire country, for the entire world.”
“That is because the data sets they are using are globally generalizable. One model that works here can work everywhere.”