During my PhD, I have been working on Evo-Devo patterns (especially the debate around the
hourglass model) in transcriptomes, with an emphasis on adaptation. I have characterized
patterns in model organisms in terms of constraints and especially in terms of positive selection.
I found that the phylotypic stage (a stage in mid-embryonic development) is an evolutionary
lockdown, with stronger purifying selection and less positive selection than other stages in
terms of the evolution of protein sequences and of regulatory elements. To study the adaptive
evolution of gene regulation during development, I have developed a machine leaning based
in silico mutagenesis approach to detect positive selection on regulatory elements.
In addition to transcriptome evolution, I have been working on the tension between precision
and stochasticity of gene expression during development. More precisely, I have shown that
expression noise follows an hourglass pattern, with lower noise at the phylotypic stage. This
pattern can be explained by stronger histone modification mediated noise control at this stage.
In addition, I propose that histone modifications contribute to mutational robustness in
regulatory elements, and thus to conserved expression levels. These results provide insight into
the role of robustness in the phenotypic and genetic patterns of evolutionary conservation in
animal development