In October of 2015, Tanjona Ramiadantsoa walked into my office at Princeton University. “They tell me you work in my country!” he said. I glanced up from the dissertation chapter looming on my computer screen to take in his stereotypical Maki-brand baobab T-shirt. “Malagasy ianao!” I said. “You are Malagasy!” And a new collaboration was born.
I wrote to you last as a Princeton doctoral student in Ecology and Evolutionary Biology, where I studied the transmission dynamics of potentially zoonotic–or human-infecting–viruses carried by Malagasy fruit bats. Disease ecologists like myself use mathematical modeling tools to understand how pathogens persist in finite host populations over time–and to predict when such pathogens are most likely to pass from one individual to another. I wrapped up that PhD a few months ago and started a postdoctoral fellowship with the Miller Institute at UC Berkeley, but I’m still chasing answers to many of the same questions as before.
Tanjona is a Madagascar-born mathematical biologist who did his PhD training under the celebrated father of metapopulation theory, Ilkka Hanski, at the University of Helsinki. Like me, Tanjona is now a postdoctoral fellow – he at the University of Wisconsin-Madison – and he studies how environmental degradation affects biodiversity loss. Conservation biologists use metapopulation theory to model how species might survive habitat fragmentation by seeking refuge in neighboring patches of still-intact habitat. Typically, species with better dispersal capacities, or the ability to move between distant habitats, are more resilient to fragmented terrain. As Tanjona will tell you, disease ecology and conservation biology are essentially the same thing, except that disease ecologists typically are interested in making their pathogens go extinct, while conservation biologists are doing their utmost to help their species avoid such a fate.
Together, Tanjona and I organize the eight-person instructor team that teaches E2M2: Ecological and Epidemiological Modeling in Madagascar, a week-long workshop in quantitative biology aimed to introduce Malagasy graduate-level students in science and public health to the use of models in their research. We teach the course at Center ValBio, a research station in spectacular Ranomafana National Park, close to the offices of one of our partners, the health-care NGO, PIVOT. In keeping with the spirit of E2M2, PIVOT uses quantitative modeling tools to evaluate the impacts of their healthcare interventions, and their research advisor, Andres Garchitorena, is one of our instructors. Our course is additionally supported by Princeton’s Center for Health and Well-Being and the Institut Pasteur of Madagascar (IPM).
“What is a model?” Tanjona asks on day one, showing a slide with an image of an attractive man in a tight sweater. The class laughs appreciatively.
Models are simplified versions of reality, and we use them to try to understand patterns in the real world. “All models are wrong,” statistician George Box, once famously said, “but some are useful.” As scientists, we know that we are largely incapable of explaining all complex processes in an ecosystem, but we hope that by building models of simplified subsets of reality, we might better understand at least a few aspects of the larger ecosystem. Tanjona and I primarily build models in the form of mathematical equations, which produce projections of interacting population densities over time – we can model populations of viruses, infected hosts, or endangered species in much the same way. And just as we hope that the sweater might look as good on us as it does on the model in Tanjona’s powerpoint, we aim for our equations to accurately recapture the past or predict the future.
It’s been a full year since I last set foot on the Eighth Continent, my longest absence from Madagascar since I started my doctoral research in 2013. The past year was exhausting and sometimes disillusioning—what with defending my dissertation, guiding my PhD chapters through peer review, and moving across the country. My Malagasy is “votsa” – rusty, like a dull knife – on arrival, and my mind is scattered. But it takes only a few breakfasts of sugary coffee and vary amin’anana (soupy rice) to feel at peace once more.
“Should I remake last year’s tutorial to match the model we went over in class today?” It is evening after a long day of teaching, and I peer at tomorrow’s schedule as I ask advice of Jessica Metcalf and Amy Wesolowski, fellow E2M2 instructors and assistant professors, respectively, at Princeton and Johns Hopkins University. “It will be more work but better teaching,” says Jess, and there is no question left in my mind. I feel rejuvenated, refreshed, and inspired to do the best I can for the students at hand. And I think to myself—this is how teaching is supposed to feel.
Next morning, I present the revised lesson on model fitting – how to adjust those equations to make the sweater fit the data a bit better – and the students are interested but also confused. “Rado would like to clarify a few points,” my longtime colleague and fellow instructor, Christian Ranaivoson, whispers to me as I wrap up the tutorial. And Rado J.L. Rakotonanahary, a scientist with IPM’s Plague Unit, steps in to explain the concept of maximum likelihood – a statistical method used to measure how closely a model replicates the data – in more sophisticated Malagasy than I am likely to ever learn to speak. I feel the pride of teacher-turned-student as I watch the class nod in comprehension.
At the end of the week, we return to Madagascar’s capital, Antananarivo, and the students present their independent research questions and model frameworks in cautious but excellent English at Institut Pasteur. I smile as I watch Soa Fy Andriamandimby, of the Virology Unit of Institut Pasteur, present her theoretical approach to rabies eradication on Madagascar Island, and Andry Ny Aina Rakotomalala, of the Department of Entomology at the University of Antananarivo, passionately describe his network model on the behavioral dynamics of invasive-native ant interactions in Madagascar—and his plans for lab-based experiments to test it.
“In a few years, I am going to be obsolete here,” I laugh to fellow instructor Fidisoa Rasambainarivo (‘Fidy’), a Malagasy veterinarian and PhD student at the University of Missouri-St. Louis. Already, three course alumni—Jean-Marius Rakotondramanga, Ornella Assimini, and Antso Raherinandrasana, respectively of IPM, IPM, and the Madagascar National Institute for Public Health— serve as mentors who teach a subset of E2M2’s curriculum. In the future, we hope to take on even more.
Fidy tells me that he is in the process of starting up a new Madagascar-based laboratory, which he calls “Mahaliana” – to spark interest. His lab’s slogan – “It always starts with a question” – captures the sentiments of E2M2 well. Critical thinking and creative, independent research ideas are more valuable than any computational skill that we might teach. Yes, programming is one component of E2M2, but the art of building a useful model is what we are really striving to convey.
We close out the week with an instructors’ meeting with Julio Rakotonirina, Antso’s supervisor at the Madagascar National Institute for Public Health, who serves as our program evaluator. Julio meets with the students both during the week and after to gather anonymous feedback for how best to improve our workshop. There are criticisms, of course, but on the whole, the students are very happy. I close my eyes and recall their singing at our final banquet, momentarily lost in the magic of Madagascar and the euphoria of a job well done.
Featured image caption:
E2M2 student, Tsilavo Razafimanantsoa, investigates a brown leaf chameleon (Brookesia superciliaris) in Ranomafana National Park. January 2018. Photo by Cara Brook.