Charles Fisher joined the Institute in November 2014. He had gotten his PhD from Harvard in 2012 and did a first postdoc at Boston University. He works mostly on theoretical ecology.
The human body hosts thousands of symbiotic microbial species making up what is called the human microbiota. Elucidating the roles of different ecological forces in structuring the human microbiota is an important step towards understanding how the microbiota affect human health. The Human Microbiome Project and similar studies have taken a step towards this goal by collecting microbiota samples from multiple body sites across a large population of different individuals. A few stylized facts about the microbiota have emerged from these studies. First, the composition of the microbiota of a body site in a particular person varies less over time than across the population. Second, the variation in species composition between body sites is much larger than the variation of a single body site across a population. These results tell us that even though a particular body site (e.g., the human gut) shares common features within a population of different individuals, every individual has a different microbiota. How, then, is it possible to extract information about the ecological forces acting on the microbiota when every sample is taken from a different environment ?
Most of Charles Fisher’s research at ENS has focused on understanding how large ecological communities vary across different environments and how to extract information about species interactions from these data. The hypothesis underlying this research is based on a classical idea in theoretical ecology that species abundances are chosen to pack species into a ‘niche space’ so as to utilize all of the resources in the environment. Specifically, he is studying a model for relative species abundances based on two simple principles : (1) the equilibrium relative abundances ensure that the species fill the entire niche volume in a way (2) that maximizes the diversity of the community. Following this idea, the relative abundances of the species can be found by solving a specific maximum entropy problem. This simple model provides a number of helpful insights into the ecology of the human microbiota. One of the most important features is that it is possible to infer the parameters of the model exactly using a technique from machine learning called Independent Component Analysis.
Currently, C. Fisher is working on applying these techniques to data obtained from the Human Microbiome Project.
C. Fisher collaborated with Thierry Mora and Aleksandra Walzcak (ENS). Other collaborations are with Pankaj Mehta and Daniel Segrè (Boston University) on related projects in theoretical ecology. C. Fisher published several papers while at the Institute [28, 29, 30].