Inside Singapore's Mosquito Factory | Yanni Yoong, Debug | Tank Talks Asia

YANNI INTV: 260518

MANISHA: ?Hi, it's Tank Talks Asia. I'm Manisha Tank at the AsiaWorks studio in Singapore, bringing to you a very special episode of the show that should coincide with World Dengue Day.

Mosquitoes. I gotta admit, I have a terrible relationship with them, and I really don't enjoy the bites, and I'm sure many of you listening don't enjoy them either and that is why my co-host and co-founder in the show, Andrew Clark, was the one that I sent into a mosquito-making factory.

ANDREW: And not just any old mosquito, Manisha. The Aedes aegypti mosquito, which is the one that's behind dengue, right?

MANISHA: And Zika.

ANDREW: And Zika, and chikungunya.

MANISHA: Now, look, we jest a little, but these are really serious diseases, aren't they? Definitely Singapore has a problem with dengue, and has been working on trying to eradicate it for some time, and in India every year, there are many cases of dengue, chikungunya, and Zika has been a problem in South America and other parts of the world, as well as here in Asia.

So how incredible to go and visit Debug, which is working on solving this problem, and I think what's interesting, it's sort of an arm of Google, and we're gonna learn a little bit more about how that works. What has a tech company got to do with mosquitoes? But you ventured into the fray.

ANDREW: That's right. I went right into this remarkable facility where they just make mosquitoes, loads of them.

ANDREW: I just wanted to stop for a second and just make sure that you know that they produce 10 million mosquitoes a week right here.

ANDREW: I have never seen so many mosquitoes in my life.

YANNI: Yes. So one of these cages contains 250,000 mosquitoes, male and female.

ANDREW: 250,000 mosquitoes inside there?

YANNI: Yes, yes, it is a big cage.

ANDREW: I'm being attacked by mosquitoes now.

MANISHA: Look at you moaning about getting bitten by mosquitoes. You got to go into the factory. That was exciting. But who was that lovely lady with you?

ANDREW: That lovely lady was Yanni, and you had the pleasure of sitting down and talking with her, didn't you?

MANISHA: Yeah. Let's get to know her a little bit better, shall we?

So who are you?

YANNI: So I'm Yanni Yoong, and I run a mosquito factory.

MANISHA: You run a mosquito factory. Do you say that at dinner parties when you meet new people?

YANNI: Yes, and I always get surprised looks from everyone.

MANISHA: Obviously, we want you to explain what that really means. Why would you wanna run a mosquito factory here in Singapore?

YANNI: Okay, so in Singapore I'm part of Debug. We're all working on ways to scale the production of mosquitoes and also the releases. So in Singapore, we have this facility that produces more than 10 million mosquitoes per week, and that is being released in different parts of Singapore, with the aim of actually bringing down the mosquito population in Singapore.

MANISHA: So what is Debug exactly?

YANNI: So Debug is a team of engineers dabbling in hardware and software, and scientists looking for ways to scale the production of mosquitoes.

And over time, we worked in different parts of the world, including in Singapore and we are an independent team under the Google umbrella, and really working on something that the rest of the Google team has not done before.

But really focusing on the use of AI, which is what, you know, our whole technology is grounded in, and using AI to actually, you know, help us with the rearing and also the releasing of the mosquito.

MANISHA: Yeah, 'cause I mean, it's really confusing for a lot of people. Because why should Google have anything to do with breeding mosquitoes?

YANNI: Yes, exactly. The main thing is we're trying to address or trying to solve dengue. So being in Singapore and being at Google at this time allows us to be able to say, "Hey, we can make a difference. We can solve problems and address real-life situations."

MANISHA: Is it working so far?

YANNI: Yes, it's working. So in the release site, it has been reported that the population has gone down very low. The risk of dengue transmission has also dropped.

MANISHA: We were lucky enough as a team to go backstage at the factory and look around. You made us all wear long pants, trousers, whatever you call them wherever you're listening in the world, and light colors. We had to wear these special..

YANNI: Jackets.

MANISHA: Debug jackets.

ANDREW: So we have to get changed?

YANNI: Yes. So grab a jacket.

ANDREW: Okay.

YANNI: Yeah, so this is a light-colored jacket, so when you wear them and you work in the factory.

ANDREW: Yeah.

YANNI: And you move from doors to doors, rooms to rooms, you will not bring out any escapees because they're visible in front of you. If I'm, for example, I'm wearing black.

ANDREW: What you get mosquitoes escaping from here?

YANNI: Yeah, sometimes they like to escape. They're very good at hitchhiking. They're great hitchhikers. So by wearing this, when you are getting through the ante rooms, which basically are small rooms in between two big rooms, so these ante rooms are for you to stop, check and this kind of jacket will help you, you know, help increase the visibility of the mosquitoes.

If it's black, if it's a mosquito hovering around here, you can't see them, but this white jacket will help to increase the visibility, and then you kill them before you leave.

ANDREW: And what do you use to kill them, your hands?

YANNI: Oh, we have our rackets. We have plenty of rackets.

ANDREW: Ooh. And one last question: Are we gonna get bitten?

YANNI: With our long sleeves and our jackets, we will not. But the thing is there could be escapees, so in this case it's better to kill them before you leave the place.

MANISHA: Just ahead, why you need to mass produce mosquitoes to make them disappear.

So how are you doing that?

YANNI: This mosquito facility, production facility, actually focuses on the usage of Wolbachia.

Wolbachia is a naturally occurring bacterium that is present in more than 60% of the insects around the world. How it's being used here is Wolbachia is injected into the eggs of Aedes aegypti, which is the main vector for diseases such as dengue, zika, chikungunya. By doing that, you're actually creating a colony of mosquitoes that you can either use it to mass produce mosquitoes or you can use it to, you know, feed it back to the colony and maintain the egg bank.

So the purpose of making these mosquitoes is, with Wolbachia mosquitoes, we have to rear them and also sort them and also release them. So what I mean by that is when you rear them, you want to be able to rear Wolbachia mosquitoes with the purpose of being able to separate them into male and female, and you do it in the pupa stage and you do it in the adult stage.

So the reason for doing this is when you have Wolbachia, male mosquitoes in this case, being released into the wild and finding a female to mate with, what happens is a Wolbachia mosquito that's the Wolbachia strain versus a female mosquitoes that doesn't have the strain, it creates this incompatible mating.

And when this happens, the eggs will not hatch, and this creates a population crash. So with each generation, the population becomes smaller and smaller.

The factory that we are currently running right now is able to produce mosquitoes at scale, so we're talking about millions of mosquitoes, and also these mosquitoes are able to be accurately sorted to male and female.

Because for this technique to work, you can only release male mosquitoes, and the interesting thing is that most people wouldn't know is male mosquitoes do not bite. So it's safe to release them. The bacterium is harmless. It's naturally occurring and it does the job of mating with females and with the eggs that will not hatch. So it does help to reduce the population of biting mosquitoes.

YANNI: We have about 250,000 males and females mating, and they produce about three million eggs.

ANDREW: Three million eggs.

YANNI: Three million eggs.

ANDREW: So there's something like three million eggs here right now or maybe more than that.

YANNI: This comes from many different cages as well.

ANDREW: Right. Okay, so there's definitely more than three million eggs here at this moment in time.

YANNI: Yes definitely. Main thing is based on our rearing protocol, this is the number of eggs that we can produce.

MANISHA: Coming up, from maps to mosquitoes, the ultimate career pivot.

What was your journey like because you were not a scientist, but then you've built a mosquito making factory and you clearly know your science on it.

YANNI: So I'm neither a scientist, I'm also not an engineer, with no background in this at all. But I did started working in Google Maps.

So on Maps I really worked on, you know, a lot of useful project in terms of helping to make maps useful for other people. And one day I came across a job that says, "Okay, we are looking for a field trial manager," so someone to run the field trials in Singapore.

And back then Debug was starting to, you know, enter Singapore and trying to work with the local government, and they needed someone to be able to run the field trial operations and, you know, the part about working with the client and getting operations running. So that's how I started, and I'm like, "Okay, this is interesting. I can do useful stuff again." You know? You're trying to really bring down the mosquito population and make sure that people don't fall sick. And I signed up for the job.

And in Singapore we started with small trials, and that was when we worked with the government. We went out into certain release site to do mosquito releases. We also made mosquitoes at a smaller scale and to demonstrate the effectiveness of our equipment rearing, sorting, and also the release mechanisms.

MANISHA: So how long ago were those initial trials?

orking with the government in: 2018

MANISHA: Okay.

YANNI: So that's been about almost 10 years in Singapore and during this period of time we started really small, very scrappy, really learning how do mosquitoes fly in a HDB setting, how do mosquitoes fly in the landed setting.

Humidity, the weather, all plays a part in how we rear the mosquitoes and how mosquitoes perform in the field, right? There's a performance rating for them.

Doing all this work allow us to kind of ramp up our production and the research that we have done behind all this, and also working with our team in the US, because people in the US don't really know how operations are run in Singapore.

So a lot of the intel information that we get from Singapore, we have to feed this back to the engineering team and the scientists, which helps us, you know, create and design programs that's useful for Singapore and within the context of Singapore, which is high-rise buildings like HDB and landed properties, which is also very unique compared to other parts of the world.

YANNI: So do you wanna take a look at the mosquitoes that we have sorted so far?

ANDREW: Okay.

YANNI: It's either male or female mosquitoes in most instances.

ANDREW: There's a picture of two mosquitoes over there.

YANNI: So one of it's a male and one of it's a female, and this is what the AI sees when they walk down the path and the photos get taken.

So, like, they have to determine in split second where to send the mosquitoes. So over here, do you think you can determine which is male and which is female?

ANDREW: Obviously to me, the dainty female, the smaller one with all of the frilly bits at the bottom on its proboscis, compared to the big, ugly male on top, right?

Male on top, female below.

YANNI: That's not the case.

ANDREW: Oh, really?

YANNI: This is the female and this is the male.

ANDREW: No way.

YANNI: Yes, it is.

ANDREW: So it's only the female that bites.

YANNI: It's only the female that bite. That's right. Male mosquitoes do not bite. They only look for sugary solution to get the fuel they need.

ANDREW: Well, I stand corrected. That's absolutely fascinating.

MANISHA: In a moment, the tiny challenge at the heart of a big project.

Something that really intrigued me was mosquitoes are tiny.

YANNI: Yes.

MANISHA: So being able to sort mosquitoes, that sort of blew my mind. And I thought how on earth do they manage to do this? So how do you make your observations? Because you have to make them on a tiny scale.

YANNI: So this is the hard work of all the engineers in the US. You know, we have developed the end-to-end technologies, starting from the rearing to the sorting and the releases of the mosquitoes. So during this time in the egg stage and the larvae stage, we focus on really maximizing the youth, right?

So make as many eggs as possible that are healthy, that you can, you know, hatch to become mosquitoes that we can use, and also rear them well enough so that we get enough pupae. So from then on, the stages turn into sorting. So at the pupae stage, we know that male pupae and female pupae are different in size. Female is slightly bigger than the male.

YANNI: So over here the pupae are male and female combined.

ANDREW: Yeah.

YANNI: As with any eggs you lay, it's 50% male, 50% female. So over here you can say about half of them are male, half of them are female.

ANDREW: But you need to separate them.

YANNI: I need to separate them, and most people would be able to say female pupae and male pupae, female is a little bit bigger than the male.

ANDREW: Right.

YANNI: So just using their size, their physical size, we are able to separate them.

ANDREW: And how do you do that?

YANNI: By eye.

ANDREW: By eye?

YANNI: I'm kidding. No, we have many thousand, millions of pupae, so we can't do that.

ANDREW: Yeah.

YANNI: So we do have robotics equipment. So there's some machinery that's able to help us sieve them.

YANNI: So we use a mechanical automation and a mechanical push to actually sort them.

MANISHA: But it's basically a sieve, right?

YANNI: It is a sieve, yes.

MANISHA: 'Cause I was thinking that must be really complicated. How do you do that? But then you showed me and it was a sieve.

And I think you showed that to Andrew as well when you did the tour.

YANNI: Correct.

ANDREW: This is an enormous sieve.

YANNI: Yeah ...

ANDREW: Doing a first separation

YANNI: That's right.

ANDREW: of the males and the females. But it's not very accurate, is it?

YANNI: At this stage we are trying to take out as many males as we can. So a male pupae, as mentioned, is smaller than female pupae.

So it has to go through a mechanical sieving process that separates the two of them based by their size. So we have two stages of sieve here. The first one sieves out the larvae. So as I mentioned, there could be some smaller larvae that hasn't quite developed, and that stage is to get rid of them.

Once the larvae, alpha larvae has gone through the sieve, what stays on top are male and female. And when they come to the next stage, and this is a male and female sieve. So when you sieve through the male and female pupae, the male pupae, being smaller in size, goes to the bottom of the sieve.

The females stays at the top. And this is where we either discard the female or we reuse them somewhere.

ANDREW: Because all you're looking for is males.

YANNI: Males, yes. So what gets sieved through are mostly male. At this point we won't say they are 100% accurate, but they will be mostly male that gets through. And then after that we'll collect them and we'll put them into our sex sorters.

MANISHA: Coming up, AI, the secret sorting weapon.

I think a lot of people are talking about AI at the moment. And I think this was one of the first examples that I've seen in real time of AI being leveraged to help us sort through something that I think human beings just couldn't do. Like, to sit there all day long and figure out what's a male mosquito, what's a female mosquito. How did you feel the first time you actually saw it online and saw it working?

YANNI: That brings us to the stage of the adult sorting because AI is being used in the entire adult sorting process. We put the male pupae into a sex sorter. So as the name suggests, it is meant to sex sort the mosquitoes based on the morphological features.

It's not possible to do it, you know, with a human being. So this allows us to do it for a long period of time with high level of accuracy and with that process, we are able to keep learning for the process and get more and more accurate.

The AI stage allows us to get through the first stage of quality check on accepting only males and the next stage is when you have the photos of all the males that have been taken, machine learning process goes through all this picture one more time to determine if there is a possibility of that mosquito, you know, not being a male.

And that case is being escalated to a human qualifier. So this is where our scientists and our engineers actually do go through the photos and determine, okay, this is okay. This is a male. It's fine. This is where I will say in terms of what the facility is all about, the magic happens.

ANDREW: We've now come into this room which is full of activity. There are guys who are washing things. There are guys with these huge trolleys, and then there are these weird machines. What on earth is going on here? What's happening?

YANNI: So this is our sex sorters. So as the name suggests, it is meant to sex sort. And in this case, we sex sort the adult mosquito. So earlier, we talk about separating the pupae by size, and once the pupae has been separated by size, and mostly the male pupae are consolidated, we bring them inside here.

So they get loaded into the sex sorters, and the main purpose of the sex sorter is allow the pupae to eclose to become an adult. And once they're an adult, we have a way to make them walk and march down in a straight line on the sorter lane, and this is where they get a photo taken.

ANDREW: So what I'm seeing here is it's like a lit up corridor.

YANNI: Yes.

ANDREW: Where the mosquitoes are walking along, and then suddenly a door opens and closes, and that's them being sorted.

YANNI: That's them being sorted. So there is a camera at the top, so there's a added process of a photo being taken. If we go by the proper procedure, the mosquito first walk down a path. Let me show it to you here. They walk down the path. The first mosquito passes through a shutter and we'll make it such that one mosquito gets through.

And once it passes through the shutter, the door closes. When it's closed, the camera will take a picture of the mosquito in that path after the shutter. When several photos are being taken, the AI here determines if they are male or female based on morphological features.

MANISHA: How many people do you have working in the factory?

YANNI: We have full-time employees, we have about 20 people. That includes our scientists, our engineers, and also our operations staff, right? So they are the ones doing the planning. They are the ones making sure that our factory is run optimally and then we also have operators that comes in on the daily basis to work in the facility and they also work outside the facility, and we're talking about 30 to 40 over people for each of this team, and they go to the field. They do the releases. They come back and in the factory, they produce the mosquitoes and a lot of the work that involves, you know, making sure the equipment is serviced properly, is cleaned, is maintained, it really lies on their shoulders.

ANDREW: So this must be the last stage.

YANNI: It is.

ANDREW: What are these?

YANNI: So these are containers of mosquitoes that has been sex-sorted. So they're male mosquitoes that has been prepared for releases. So they go into a room, they get staged overnight, and in the morning we have staff coming in to bring them out for releases.

They get put into release vans, where there's a driver that drives them around, and once they put it into the release mechanism, they get released evenly, distributed. And over a large area, you just need one driver carrying, carrying them and releasing them in a very automated fashion.

ANDREW: Wow. So the male mosquitoes only inside these what look like magazines, really. Into a car, and then as the car drives around, it releases them into the wild.

YANNI: That's right.

MANISHA: Still to come, can this Singapore-built solution really protect millions?

It's probably a good thing that this is happening in Singapore because if you look at Asia, dengue is a problem in a number of Asian countries. Am I right in thinking that most of the world's burden when it comes to dengue is focused or concentrated in this part of the world?

YANNI: Yes, as of today, about four billion of the world's population is at risk of dengue, and 70% of that lies in Asia.

MANISHA: Wow, okay.

YANNI: So which means to say that we are in a very good place. Singapore is located in the middle of Southeast Asia, and we have the strategic advantage to be able to, you know, innovate out of the R&D hub that we have in Singapore.

And also to extend, you know, the science knowledge as well as all the engineering capabilities that we have in order to bring, to extend the efforts that we have in mind right now. There’s the suppression technique that works in Singapore. It's just, you know, finding the way to extend them and also to leverage the kind of the know how that we currently have.

MANISHA: Have you had much interest from people in the region or even around the world about the project you're working on?

YANNI: Yes, so we do have a fair amount of interest from different parties, you know, be it NGOs or private companies. I think it's also a good time because right now our facility is equipped, you know, with different ways, you know, knowledge of the scientists and also the efforts of our engineers to develop solutions that could be customized and tailored to other Asian or Southeast Asian countries.

MANISHA: I think the world knows that Singapore is a fairly wealthy country and it costs money to build a mosquito factory and then to actually run a project like this.

But unfortunately, Zika, chikungunya, dengue, these affect countries which aren't always flush with cash. How's it gonna work for countries that can't afford it?

YANNI: So in Singapore we use the suppression method, which is basically using Wolbachia and releasing only male. But there are also other methods that can help to tackle the problem of spread of diseases like dengue.

So the other method's called population replacement and instead of releasing male only, which is needed if you only want to reduce the population, but this method, population replacement, is where you release both male and female in order to take over the wild population and only have Wolbachia males and female living in the environment that human beings are staying in.

Because mosquitoes with Wolbachia does have a natural blocker for the spread of diseases such as dengue. There’s another method and that method can be customized, that method can be cost-effective depending on the type of market we work with, and we can find ways to deliver this kind of alternative solution to countries in the Southeast Asia where 70% of the dengue burden lays in.

MANISHA: So you said earlier that things are working. The suppression is happening. Can people in Singapore look forward to no dengue anymore?

YANNI: The key is to be able to release in the treatment sites over time. Being a small island with busy ports and so many people flying in and out. It is not so easy to just say, "Hey, no more mosquitoes come in." Because mosquitoes do come in.

MANISHA: Yeah.

YANNI: They could be in cargo. They could be in, you know, plane rides. They could come in. This method that we are deploying in Singapore, it is to reduce the population. So it doesn't mean that you can stop mosquitoes from coming in because mosquito can still come in, so there is a need to continuously maintain the area by having mosquitoes being released. And it could be lesser, but so long, you know, it does a good job of maintaining over time.

So if you were to cut it, it means to say that new mosquitoes could migrate to the area that was previously mosquito-free.

MANISHA: Yeah. I think a lot of people have assumptions about mosquitoes. But there are actually so many breeds, right? And they don't all transmit disease.

YANNI: No.

MANISHA: Specifically you're targeting, as you said before, the Aedes aegypti mosquito.

YANNI: Yes.

MANISHA: But that doesn't mean that there aren't others.

YANNI: Aedes aegypti is the key vector that spreads dengue, Zika, and chikungunya, which is primarily the key problems that all of us have. So by taking out Aedes aegypti, we obviously will make people not fall sick so easily and prevent biting nuisance as well. Doesn't mean that all the other mosquitoes are not there anymore.

MANISHA: I mean, one of the things that's troubling a lot of people is obviously climate change.

I think we're already beginning to see it, that climate change means there could be more insects that we're dealing with. Or we could see the spread of diseases. We've already had the WHO sort of warn and talk about these sorts of issues.

Do you think your job's gonna become even harder? You're gonna have to double up as the world gets hotter?

YANNI: Climate change and urbanization will cause issues, right? And especially in Southeast Asia where it's tropical region. So we continue to see issues like this, and we hope that we can use, you know, Wolbachia as a suppression method, and also even replacement method for other countries in order for us to address the situation, whether is it to replace or is it to suppress.

MANISHA: Yeah. And I think there are some people who might say, "We have no right to interfere, why are we interfering with nature?" What would you say to those people?

YANNI: I think in this case, Aedes aegypti is considered a vector that spreads diseases like dengue and Zika, which affects a lot of human beings. So in this case, we are really looking at a specific target, in this case it's Aedes aegypti, in order to create less, you know, issues, health issues, among residents in Singapore and even in the region as well.

So that is something that is quite important. There are still a lot of mosquitoes everywhere. There are different types of mosquitoes everywhere.

MANISHA: You're not trying to eradicate all mosquitoes.

YANNI: No. We're not doing that.

MANISHA: And I think what's also interesting, right, is you're using a bacteria within the mosquito life cycle to help tackle a difficult problem. It's not, and that, I would suppose, is way better than using pesticides or the other means of dealing with dengue and Zika right now.

YANNI: That's right, it's naturally occurring, so which means to say it's harmless to human beings.

And obviously it's also not as bad as like your pesticide or even other methods that is considered not ideal. So this is still one of the better methods to use in order to carry out suppression.

MANISHA: Okay. Yanni, what's your big dream? Where do you think, or where do you hope you'll get to, say, five, 10 years from now?

YANNI: I think we hope to protect the residents of Singapore, and also to extend the method to other parts of the region in Asia, and to help tackle and address dengue and the issues of vector-borne diseases.

MANISHA: And in the meantime, you're gonna excite all your dinner guests by telling them that you breed mosquitoes for a living.

YANNI: Yes, yes. Always, always a great pitch to start a conversation.

MANISHA: Yanni, thank you so much. It's been great talking to you. And good luck. It sounds like an incredible project.

YANNI: Yes. It is great to know that the work that you do can help to save lives. I think at some point it saves life, and it also helps to make people fall sick less often, so that's great.

MANISHA: And ironically, I have a mosquito bite. I didn't get it here, though.

YANNI: Good to know. At least not in this room.

MANISHA: I think that's quite something, isn't it, Andrew, to be able to say that your job is something that leads to saving lives, right?

ANDREW: I think that one thing which is really interesting to me is why Singapore is clearly the best place to make this happen. The fact that so quickly and so efficiently from a facility like that, that produces 10 million mosquitoes a week, you can very quickly and easily deploy mosquitoes around the region to help solve and tackle that challenge.

MANISHA: Yeah. But unfortunately, and Yanni did point this out to me didn't she, there are more than 100 different types of mosquitoes. So you still could get bitten, so don't listen to this episode and think that you can go out without wearing your mosquito repellent. I strongly advise it. I take it everywhere I go. But anyway, there are gonna be loads of notes in our show notes, 'cause I'm sure there are gonna be lots of questions coming from this one. And what does everybody need to do?

ANDREW: Everybody needs to like, everybody needs to subscribe. But actually, one thing that we should ask our listener to do is also to leave us a review and give us a rating, because I hear that that really helps in terms of getting our show out to people like you who like to listen and watch it.

MANISHA: Indeed. And I have one more request, which is send us your ideas.

ANDREW: That's such a good idea.

MANISHA: Because I'm always hearing from people who suggest we should do a story on this or that. It's like, okay, write in to us, because we'll definitely pick up on it.

ANDREW: Absolutely.

MANISHA: And with that, thank you all for joining us for this episode, coming to you from the AsiaWorks studio in Singapore. We're looking forward to catching you again next time.

Tank Talks Asia is an AsiaWorks production and the views and opinions shared by our guests are their own.