I am currently a postdoc with Eduardo Rocha (Microbial Evolutionary Genomics) at Institut Pasteur, working on bacterial recombination in complex biotic environments using metagenomic and bioinformatic approaches.
Before this I was a postdoc with Sebastian Bonhoeffer (Theoretical Biology) at ETH Zurich, working mainly on stress-response, evolvability and robustness in bacteria (read our recent publication here).
And before I was a PhD student with François Taddei and Dusan Misevic (INSERM U1001). My PhD focused on the links between second order selection and evolution of cooperation, using both microbial and in silico systems (some of our work here and here).
We use metagenomic datasets to extract quantitative properties of the biotic environment of bacterial species. The goal is to investigate how different lifestyles and biotic environments are linked with the use of recombination and horizontal transfer, as well as with the use of competition machineries (eg toxin or antibiotic production).
Stress-induced mutagenesis has been a major paradigm shift in the past decades: it postulates that as an answer to stress, bacteria increase their genome-wide mutation rate. This has been interpreted as a mechanism providing “adaptation on demand” (increased evolvability under stress, increasing the chances that a descendant is able to face the stress).
In our recent paper (Frenoy & Bonhoeffer, 2018, PLoS Biology), we challenge this view by showing that (1) current methods lead to a systematic over-estimation of mutation rate under bactericidal stress, and (2) a stress that increases mutation rate does not always increases evolvability, in particular when it decreases population size.
To do so, we developped an experimental method to measure death rate in bacterial population using plasmid segregation, a computational method to estimate mutation rate when there is death, and a measure of evolvability that encompasses mutation rate, population size, and population turnover.
Because genes coding for cooperation (here public good secretion) face very different selection pressures than more classical genes coding for private traits (affecting only the individual bearing them), we wondered whether they would somehow evolve different genetic properties.
To answer this question, we adapted the Aevol platform to the study of cooperation by implementing a spatial structure and the potential to secrete a public good. Aevol is an individual-based model that has a bacterial-inspired genomic layer and is has been used to study second-order selection pressures acting on genome organization.
We found (Frénoy et al, 2013, PLoS Computational Biology) that genes related to cooperation (coding for secretion of a public good) tend to spontaneously form operons (using the same promoters and terminators) and overlap (using the same base pairs but in different reading frames) with “metabolic” (only contributing to the focal individual's private fitness in our vocabulary) genes. A large part of “cheating” (decreasing secretion) mutations are thus also impacting “private” genes, causing a drop in fitness and the mutation being wiped out by selection.
We interpret this as an example of evolvability suppression (evolution of a trait constraining futur evolution). Several recent studies show the potential relevance of this kind of second order selection pressures on cooperation in microbial world (Foster et al, 2004, Nature and Dandekar et al, 2012, Science) and beyond (Altenberg, 2005, Artificial Life).
Free E. coli and Salmonella knowledge: ecosalgenes.frenoy.eu, a modern online version of the list of genes that can be selected and counter-selected in these organisms, with links to Ecocyc and to full-text references. Useful for molecular cloning and microbial genetics, may become useful for system biology now that it is accessible in an easily parsable form!
Automatically remove these annoying watermarks that academic publishers insert in PDF: with this small python script, comments and suggestions welcome!
Add links to sci-hub in the references of an article in PDF format, using this Python program.