Section: Research Program

Research axis 4: EvoEvo

The EvoEvo project was a highly interdisciplinary project ranging from wet experiments (experimental evolution) to software application (evolving a musical personal companion).

Computational experiments

To help understand the results of the wet experiments conducted by our partners, we designed several computational experiments that used the models developed by the team (see Research axis 2 above).

Dynamics of innovation in viral strains. Viral strains show a surprising evolutionary dynamics in which long periods of stasis are interrupted by short evolutionary bursts. To understand this dynamics, we launched a very large scale in silico experiment in the Aevol software [40], [60].

Dynamics of mutator strains. The mutation rate of bacterial strains is known to adapt to the evolutionary conditions through the emergence of “mutator strains” that have a 10 to 100 fold raised mutation rate compared to the wild-type. Mutator strains are supposed to raise in specific conditions such as stress or environmental change. We studied this situation in the Aevol software and showed the importance of the reorganization of the genome structure in this phenomenon [76].

Diversification in seasonal environments. One of the targets of the EvoEvo project was the evolution of open-endedness. As a first step, we provided a careful definition of open-endedness [34], [85] and developed a new model, Evo2Sim (http://www.evoevo.eu/evo2sim/) [72], that contains all the ingredients to enable the emergence of structured populations. We then used it to simulate the evolution of populations in batch cultures and continuous conditions [73], [74], [75].

Bioinspired computation

Once EvoEvo strategies were identified in vivo an in silico, the final objective of the EvoEvo project was to transfer them into the ICT domain by developing new evolutionary metaheuristics and applying them to real problems.

Evolutionary-based subspace clustering. One on the main outcomes of the in silico models was the importance of chromosomal rearrangements (and of their consequences on the genome organization) for the evolutionary process. Hence, we proposed to use these properties to develop new evolutionary algorithms specifically dedicated to the subspace clustering task [69], [68]. We also developed the first evolutionary algorithm dedicated to the subspace clustering of datastreams.

EvoMove: A musical personal companion. To validate our evolutionary-based subspace clustering algorithm, we used it to develop EvoMove, a musical personal companion to be used by dancers and musicians [12], [24]. EvoMove has been tested by professional dancers (http://www.evoevo.eu/evomove-working-session/) and has been presented to the public in the context of the “Meute” (“Herd”) performance that has been played eight times between February and July 2017.

Current Objectives

Research axis 4 is relatively new in the team since it started with the EvoEvo European project. Hence many of the activities of this research axis are still fully active. We will carry on the cooperation on the innovative dynamics of viral strains, with the objective to study whether the dynamics we observed in viruses is also observed in other kinds of genomes. We also will investigate further the role of large scale events (chromosomal rearrangements, recombination...) on the macro-evolutionary dynamics and on the interaction between micro- and macro-evolution. Finally, we will carry on our work on evolutionary-inspired subspace clustering algorithms and on the evolutionary metaphor applied to the development of new ICT technologies with a specific focus on the integration of temporal data.