Stadler uses stochastic processes to study rates of change: How quickly do changes occur in the genetic information of humans, animals, bacteria and viruses? And how quickly are these changes inherited by offspring – or even transferred to other species?
Sometimes her work involves changes in the genome over very long periods of time. For example, working with New Zealand researchers she was able to show that the first penguin species emerged 12.5 million years ago. But when it comes to medical research and dealing with epidemics, it is the rapid changes that are particularly revealing: flu viruses mutate at such a rapid pace that different patients are infected with different mutated viruses even in epidemics that last just a few months. Bacteria in livestock farming can mutate within a matter of weeks or months to avoid the effects of antibiotics. But how fast do they pass on this resistance to the next generation or to humans?
“Access to constantly updated genetic data allows us to determine how quickly a mutant virus or resistant bacterium spreads through a city or region, and draw conclusions on the risk of infection for humans,” says Stadler. It is also possible to track a pathogen’s route of transmission: the greater the similarities between the genetic information of pathogens from two different people, the more directly the pathogen was transmitted between them.
Sparse data, smart models
During an epidemic, however, the routes of transmission are often so labyrinthine and random that Stadler has to get by with comparatively little sequencing data. Information on the genetic sequence of a pathogen generally comes from only a few patients or animals, and only at a single point in time during each infection so the situation is very much one of “sparse data” rather than “big data”. “We’re always working with incomplete snapshots,” she says, comparing it to an attempt to chronicle the history of a whole city based on just one photograph.
Like a detective, Stadler uses her scientific toolbox and creativity to extract meaning from the data. She is assisted by researchers at the University Hospital Basel. One project they work on together is antibioticresistant E. coli bacteria. Here she predicts the spread of the bacteria by collating the hospital’s patient data with data from livestock and agricultural production as well as data from wastewater. Among other things, she hopes that the genetic sequence data will provide insights into what happens when resistant bacteria emerge in livestock farming
To overcome incomplete data, mathematical models must be based on assumptions that are in line with the current state of knowledge in biology and medical science, says Stadler. Otherwise, the results will be inaccurate – as evidenced by the 2014 Ebola epidemic.
Tanja Stadler has another aspiration she would like to fulfil. Currently, she designs a separate model for each individual case: one for influenza, one for E. coli, one for penguins, and so on. But just like the members of a family, these models share certain characteristics. Based on these commonalities – and because “I can say for each model which questions it can’t answer” – Stadler is hoping to derive a “super model” that can be applied equally to viruses, bacteria, animals and humans. “That’s something I enjoy doing!”
