Merging computer science and ecology to better understand infectious disease

Identifying and understanding zoonotic diseases—infections that are spread between people and animals—can help predict future emergence risks, particularly in vulnerable populations. Combined with computational and data-based models, this research allows us to study potential diseases in the hopes of preventing outbreaks.

On this episode of Science Talks, Amy Barber is joined by Dr. Liliana Salvador, assistant professor of animal and comparative biomedical sciences, at the University of Arizona College of Agriculture, Life & Environmental Sciences. Dr. Salvador combines ecological, genomic, epidemiological, and movement data with computational and mathematical models to understand zoonotic diseases in wildlife, livestock, and humans. 

This interview had been edited for length and clarity.

If you could only listen to one genre of music forever, what would it be?

I like gypsy jazz. I would have to say that along with alternative indie rock infused with world music. 

I was very lucky on my last trip to France; I went to a bar in Paris, and they were playing and I felt I could stay there forever. 

What is your go-to snack purchase for traveling?

It used to be apples, but now I’m allergic to them. Now I would say smoothies or pretzels.

If you could make any mythical creature come to life, what would it be?

I like all mythical creatures: dragons, unicorns, fairies, vampires, werewolves, etc. 

But since I love The Mandalorian, I would say the mythosaur.

What made you choose this career over everything? What was your ‘aha’ moment?

I have a very atypical scientific career. 

I was originally a computer scientist, and then I switched to computational biology and then ecology. Initially, I wanted to be a vet, but that didn't happen. I went into computer science and fell in love with it. 

But then, after the end of my master's, I felt that something was missing. I wanted to do something more applied, so I applied for a computational biology PhD program, which brought me abroad to the United States. During my PhD, I focused on animal movement and started thinking about ecology and diseases. And that's what then brought me to study infectious diseases at the University of Glasgow. I thought animals move, they can spread diseases, and I want to know more about it. It was a good combination of applying what I knew from computer science, the computational methods, and the theory behind it, and then linking it to ecology, movement, and how disease spreads. 

It was a bit of a wavy path but it turned out to be exactly what I wanted to do. I have to say that it was a bit slower, because every time you change direction, you have to learn a new field. It is much slower than if you just have a straight path. But for me, it is very rewarding, because now I'm studying exactly what I want to study. It's very interdisciplinary, which is what I like.

What was it like when you earned your degrees in Portugal? Are you from Portugal?

I’m from Portugal, born and raised. I did my undergraduate and master's degrees at the University of Porto, a beautiful city. It was great to study there because there was a big set of student and academic culture mixed with the traditions of the country. 

I'm very blessed to have been raised there, and also studied there. We were very united, the cohorts, because we found it so challenging. There was a lot of collaboration between students and it was fun. When I moved to Lisbon to start my PhD, it was a bigger and more competitive institution, and I wasn’t used to that competitive environment. Then when I moved abroad, I realized, wow, this is what competition is. Not only  amongst peers, but also in terms of resources, publications, and so on.

How would you compare your experiences in Portugal to your experiences in the United States? 

It is more competitive in the United States. However, because there are more resources, there are more opportunities. In the end, I think it is competitive both ways. It also has to do with the field. 

At the time I was focusing on theoretical computer science. There's not a lot of money involved in terms of output. But once you start working more in fields related to human or animal health, with the possibility of collaboration with industry, those fields are more competitive for funding. Those fields are also wanting to be the first ones to discover a drug for a specific disease, so that adds to the competition. 

What was it that brought you to the University of Arizona? And more specifically the BIO5 Institute?

I was previously at the University of Georgia as an assistant professor with a joint appointment between the Department of Infectious Diseases and the Institute of Informatics. 

My partner was also there, and we were both thinking that it was time to go to other places, and we both got an opportunity to come to the University of Arizona. He first got the opportunity to come here and then I came as a spousal hire. BIO5 was a great fit because it is a very interdisciplinary environment. I requested if there was space for me here, at least temporarily for now, and I love it. I'm still very new here, but I hope to talk with more people, and perhaps even extend my line of research and work with colleagues.

Could you expand on exactly what you’re researching and what goes along with it?

I'm here at BIO5, but my home department is the School of Animal and Comparative Biomedical Sciences, part of the College of Agriculture, Life & Environmental Sciences. 

That is a perfect place for me in terms of a department because what I do is comparative biomedical science. I focus on zoonotic pathogens, mostly on zoonotic bacterial pathogens that cause infectious diseases in animals, but also in humans, and try to understand how these pathogens jump the species barriers and transmit across space. 

There’s a lot of debate if, for example, SARS-CoV-2 has a zoonotic origin or if it came from a lab leak. Scientific evidence shows that it comes from a zoonotic source. 

Can you tell us how you create the models you work on?

I'm trained as a computer scientist, but during my PhD my adviser, a mathematical ecologist, exposed me to many mathematical models. 

Because of the focus on ecology, I also learned statistical models, something that often isn’t learned during a computer science degree. During my PhD, I expanded my knowledge of modeling in general. But there are different types of models that we develop in the lab, mostly individual-based or agent-based models that are very computationally intensive. 

For example, if I want to study the transmission of an airborne pathogen disease from one individual to another individual, I put one person as an agent, and this person will perform certain tasks, and if it gets close to another person, it will get transmitted. That is an example of what an individual-based model could be. Models have simple rules of interaction that are usually very computationally intensive, depending on the number of individuals. If we want to simulate a disease spread in a city, you can imagine that that will be extremely intensive in terms of computational power. So, we will make many assumptions and minimize our sample size. 

Then, we also develop mathematical models, which are focused on continuous-time differential equations. We develop an equation, solve that equation, and then simulate it. These types of models are very interesting to learn. For example, at the population level, we can understand how a disease will spread in the population, depending on certain management or control strategies. 

We use a lot of statistical models as well, and also machine learning models, including where everybody's talking about: AI. Now, as a computer scientist, I think AI has been around for a long, long time. It's nice to see these applications to biological problems and data science problems. 

Are there particular diseases you work on?

I focus mostly on animal tuberculosis (TB), which is a zoonotic disease, since humans can also get the animal strain of TB. I also have worked on other bacterial pathogens like salmonella, or leptospira, and these are bacterial pathogens that can infect animals and in turn infect humans while living and persisting in the environment.

Can you share any exciting moments that you've had recently in your lab or your research as a whole?

There are two main exciting moments that I would like to share. 

One has been when I was still a postdoctoral researcher, and I had the opportunity to do science but with impact. We were working with the government and looking at what would be the effect of certain surveillance strategies to control a disease. The government was very interested in using the surveillance strategies that we designed and wanted to possibly implement them at the regional or national level. 

In the United States, we were studying the role of one species of elk to see if they could be the new animal reservoir for bovine tuberculosis in Michigan. In Michigan, whitetail deer maintain the disease of bovine tuberculosis and sporadically infect cattle. Government officials wanted to know if infected elk could be a new species that could spread the disease to cattle herds. 

It was important to work closely with the Michigan Department of Natural Resources and with the United States Department of Agriculture to understand that problem. In the end, we saw that elk were not playing a role in the maintenance and the spread of the disease. 

Nowadays, what is more exciting is to see my lab members succeed. Last year, I had three students who graduated with their PhD.  It was rewarding to see them finish and find great jobs. One had received an Intelligence Community Postdoctoral Research Fellowship, one went to St. Jude's Children's Research Hospital as a senior bioinformatician, and the other one is here at the U of A as a postdoctoral researcher

What are some of the main research goals for you?

My main research goal is to study bovine tuberculosis. We can call it animal bovine tuberculosis. If it infects cattle, it’s bovine tuberculosis, and if it infects wildlife like deer, elk, badgers, or any mammal, it’s animal tuberculosis. 

This disease can also infect humans. I want to understand the ecological, evolutionary, and behavioral mechanisms for the transmission of this disease between animals, but also between animals and people. Tuberculosis is still the number one killer in the world, even though it is a preventable disease. We have a vaccine for it, but every year worldwide 10 million people fall ill and 1.5 million people die. 

There is a percentage of that, that is due to zoonotic tuberculosis. It's not the same species, but these two species are 99.9% genomically identical. It is very difficult to diagnose zoonotic tuberculosis in humans because the symptoms are very similar. If someone goes to a hospital or a clinic, it is very likely they will be misdiagnosed. They don't test for a specific species because it is all part of what we call the mycobacterium tuberculosis complex. 

Zoonotic tuberculosis is mostly caused by what we call mycobacterium bovis, but in humans, it is mostly caused by mycobacterium tuberculosis, but they are all part of the same complex. The symptoms are very similar, however, the treatment is different because mycobacterium bovis is resistant to one of the antibiotics usually given as a treatment. If the hospitals and clinical settings don't do a species test, they will follow the clinical routine of providing four antibiotics for the treatment of TB. However, if it is caused by mycobacterium bovis, this person will not recover, because this bacteria is resistant to that antibiotic. 

The other issue with zoonotic TB is that humans can not only become infected by close contact with infected cattle, but also through infected milk and dairy products. Contaminated dairy products will not always cause the typical lesions on the lungs. So X-rays will show no lesions, and clinics will assume it’s something other than zoonotic TB, delaying the correct treatment.

Because of these diagnostic issues, we believe that we don't know what the true burden of zoonotic TB is in the worldwide population. Especially in countries where there is no pasteurization of milk, and people drink milk directly from the cows. There are also at-risk populations that work at slaughterhouses or working farms who are in direct contact with animals. Animal TB and bovine TB are endemic in many countries. 

That is the main goal of our lab: to understand the transmission of animal TB across different animal species, as well as the human population. We also want to know if a human can affect another human if it isn't infected with that species of the tuberculosis complex. We still don't know if a person who is not previously immunocompromised can receive or get infected from another person that has mycobacterium bovis.

Can you tell us why the goals are important in the long run?

We have the goal of controlling tuberculosis by 2035. 

But it is impossible to control TB by 2035 or even later if we don't know how many of these 1.5 million people that die every year are due to zoonotic TB. it is likely a small percentage, but we just don’t know because of all of the misdiagnosis that happens. 

You have to get it to where it’s more commonplace for people to test for the tuberculosis species up front.

Exactly. They need to test for resistance upfront. And this is something that is improving under the One Health approach, where animal and human clinicians are talking with each other. 

I have been meeting with human clinicians who are more interested in zoonotic TB. They are aware that it's not only human TB that can happen to humans, but the animal species can also infect humans. They are asking their patients more questions: What is your occupation? Do you work on a farm? Do you work in a slaughterhouse? Do you eat and drink pasteurized milk? Are you in contact with wildlife? For example, there were several deer hunters in Michigan that got infected after hunting deer. The deer was infected and by dressing it, they cut their finger and were infected with TB. That is also another way of transmission. If we have that information upfront, then we can identify these different risk factors. 

Can you give us a typical day for you? How does a researcher go about their tasks?

I start the day by walking my dogs. Then, it depends if I am teaching, since the summer and fall will be more research-oriented. I plan the week before, and every night I always prepare mentally for the meetings. I am very hands-on with people who work with me. We have meetings every day to discuss the different tasks that we need to do together. 

Does it take a lot of time planning the models?

I'm working now with a new postdoctoral researcher, Sarita Bugalia and she's a mathematical epidemiologist. Together, we are developing a model to understand bovine tuberculosis transmission at the wildlife livestock and human interface, as well as the environment because the bacteria can live in the environment. 

We have been discussing the different routes of transmission and different parameters. For example, if we account for whether an infected human will be transmitted to the environment or if the infected human can be infected from the environment. For each one of these different compartments, wildlife, livestock, humans, environment, we are checking the arrows back and forth, each one of the directions if we need to account for that or not. We already have data in the literature to parameterize these models, but it can take quite a bit of time before we start.

You have offered courses in the past at the University of Arizona, how do they relate to your research?

This past semester, I taught medical and molecular biology. My research focuses mostly on bacteria, but it does have a lot of connections to transmission routes, pathogenesis, and control strategies. It was interesting to make this parallel between viruses and bacteria. They are both global but every pathogen is different. You have to ask certain questions: what is the source? Where does it go? What is pathogenesis? And then what is the evolution of this pathogen? Independently if it is a virus, bacteria, or parasite, we try to understand these fundamental questions. It was interesting to see those parallels. 

I've been learning a lot, and I'm sure they will learn even more about viruses. It was interesting to see how interested the students were. They asked a lot of questions, and they wanted to understand more and more so they could relate to it. When I talked about control measures, they were like, oh, yes, we had to quarantine this many days, we had to wash our hands, we had to do this, we had to do that, and so on. Because they related to it, it was a fun course to teach.

Do you have a mentor that has impacted your life?

I have a few that I have been very lucky to have. With the academic career, you go where the research and jobs are. I’ve lived in many different countries, and in many of them, I found people that I can relate to and be inspired by. My PhD advisor, Professor Simon Levin, has been a key figure in my scientific development and in my life. We are in touch frequently and, even at a personal level, I still contact him to tell him about good things and bad things. He has been a person who has mentored me from the beginning since we started working together until now. 

There are other people that have been very influential in my career, including Katie Hemsson, Dan Hayden, Jessica Metcalfe. I have been very lucky to have my postdoctoral advisor, Rowland Kao. I've been lucky with people that I have found since this is a challenging career. It’s important to find people who inspire you and help you think deeply about problems,andI'm very thankful to them for supporting me all of these years.

It sounds like the way your lab is structured, you are a mentor to several people.

I spent a lot of time mentoring my students, either undergraduate or graduate students and postdocs, because of all the amazing experiences I had. I felt the impact of wonderful people, and it helped me progress my career, and as a person. I want to give back to the community, be the best mentor I can be, and provide opportunities for them to succeed in the future. I always tell them, I'm here for them for whatever they need, but they also have to work hard.

Do you have anything else related to your research you’d like to share with us?

Not specifically about my research, but to any student who is listening to this podcast, I think it’s so important to find a topic they care about. Sometimes it’s not exactly where the money is, but at the end of the day, passion is worth it.

What’s next for you?

It’s been a year already, and it feels like just a few months. I’m still getting to know a lot of people here at the University of Arizona and BIO5, so next steps are to establish my lab and collaborations because I would like to stay at the U of A.

What is your reason? Why do you do what you do?

I want to make a difference. That was the main motivation for changing from computer science to biology.

Everybody was telling me, ‘You're going to have a much lower salary, you might not even have a job.’ Because at the time, computer science was a very desirable and well paid profession. Biology at the time, especially in Portugal, had limited options and a much lower salary. But I felt that I wanted to better understand our world, environment, and biodiversity. I wasn't sure what I wanted to do, but I knew it had to do with animals and ecology. Then, I ended up studying animal movement and was introduced to the world of infectious diseases. I could combine both, that's why I wanted that. I ended up trained as a computer scientist, which provided me with the tools to develop models to study disease dynamics, movement patterns, and behavioral patterns. 

I think it was just a circle that's closed. In the end, it all worked out. 

Watch Science Talks on YouTube