Stochastic models for the inference of life evolution (SMILE)
Principal Investigator: Amaury Lambert, professeur à l'École normale supérieure & Guillaume Achaz, professeur à l'université Paris-Cité
The patterns of biodiversity that we observe today, are the result of the historical interplay of complex evolutionary processes operating at all scales of time (from mutation to speciation), space (from the local displacement of single cells to the migration of populations) and taxonomy (from individual births and deaths to adaptive radiation). Depending on the scale on which they are observed, these processes can appear as continuous (accumulation of genetic variation across several generations) or discontinuous (mass extinctions on geological timescales), idiosyncratic (founder event in domestication) or generic (selective sweeps in different local populations), predictable (local extinction in a resource-depleted habitat) or unpredictable (phenotypic effects of new structural variants), reproducible (experimental evolution of short-lived organisms) or not reproducible (evolutionary transition, pandemic).
The SMILE (Stochastic Models for the Inference of Life Evolution) group aims at uncovering these diverse processes from the patterns they generate on genomic/phenotypic/taxonomic diversity by devising and studying innovative probabilistic models of evolution.
Evolution models can be broadly divided into: (a) top-down approaches, used in phylogenetics and comparative systematics, where macro-evolutionary processes are described by simple stochastic models parameterized by abstract quantities (e.g., speciation rate, infinitesimal variance of phenotypic evolution); (b) bottom-up approaches, used in population genetics and ecology, where evolution emerges from its microscopic components, namely individuals or their genes, characterized by measurable traits (e.g., fecundity, survival probability or mutation rate).
We combine, relate and apply these two approaches to four main biological questions:
Conservation genomics. We develop original population genomic models and methods to infer recent demographic variation from a handful of genomes sampled from the same population or species. Our goal is to propose scalable tools to rigorously assess the conservation status of poorly known taxa in a systematic and unbiased manner.
Evolution of human pathogens. We tailor models to explore how pathogens evolve under various selection pressures coming from their biotic (e.g., host immune cells, competing variants) or abiotic (e.g., drug) environments. This includes studying the epidemiology of SARS-CoV-2, HIV and E. coli, deciphering the origin of SARS-CoV-2, mapping antibiotic resistance mutations.
Diversification of species and molecules. We investigate individual-based models of speciation, models for the joint evolution of genomes and species, and genomic methods for species delimitation. Our goal is to understand the relation between genetic diversity and species diversity, by modeling the out-of-equilibrium dynamics where gene flow between populations counteracts genomic divergence until the latter eventually wins, resulting in reproductive isolation and speciation.
Development and somatic evolution: Although development is generally viewed as a deterministic program, we explore approaches in which the development of a multicellular organism is modeled as an evolutionary process where cell lineages divide, mutate and compete in a complex interplay between natural selection within the organism and between organisms.