My team focuses on quantitative and mechanistic approaches to evolution, from the origins of life, early major transitions, to evolutionary constraints in extent microorganisms. Central to our approach is the understanding of interaction (or reaction) networks, in vitro and in vivo.
In practice, we use laboratory evolution and high-throughput screening to establish adaptive landscapes (genotype-phenotype-fitness relationships). This is made possible by developments in microfluidics and bar-coded sequencing. Interpreting these experiments involves intensive data analysis, machine learning, mechanistic modelling of biochemical reaction networks, and modelling of evolutionary dynamics.
What are the conditions for self-reproduction?
Autocatalysis is the way by which chemical reaction network can achieve self-reproduction. Autocatalysis is known to occur in a number of experimental systems but we still lack a general approach. We have recently shown the existence of stoichiometric motifs universal to all autocatalytic reaction networks. From this basis, we aim to systematically characterize autocatalysis in chemistry, thus uncover the diversity natural and artificial metabolisms.
Another aspect of our research is to understand sequence-function relationships in catalytic RNAs (called ribozymes) capable of self-reproduction. This is achieved by back-and-forth iterations between machine learning of sequence conservation patterns and experimental screening of catalysts. Knowing the neutral space of catalytic functions in RNAs informs us on the probability of appearance of and their evolvability.
How can evolution emerge in physical-chemical systems?
We study how evolution could start in in vitro self-replication reaction networks made of RNA. By coupling sequencing and microfluidics, we address questions about the emergence of Darwinian properties (variation, reproduction with heredity, selection). Ultimately, we would like to understand how genetic information could gradually emerge from rudimentary forms of evolution.
Can we predict the evolution of gene networks?
We develop strategies to impose combined genetic or drug perturbations, and measure the resulting phenotypes. Our aim is to understand multi-factorial interactions in terms of network wiring. Our goal is to identify evolutionary constraints and ultimately test their predictive power.