Modelling the physico-chemical emergence of natural selection – EmergeNS

The origin of life necessarily implies the emergence of Darwinian dynamics in a non-biological world. To explain this transition, the first hypothesis, commonly referred to as "gene first," suggests the spontaneous emergence of objects capable of reproducing while undergoing mutations that generate heritable fitness variance, such as self-replicating RNAs. Against this perspective, the "metabolism first" hypothesis is generally opposed, emphasizing the improbability of the former and the necessity for a highly developed chemistry to give rise to objects and properties as complex as these. More realistic from a physico-chemical standpoint, this second proposition, however, faces another major difficulty: by focusing on purely chemical processes, it deprives itself of the immense explanatory power of Darwinian dynamics, which could facilitate the emergence of complexity. According to our hypothesis, developed within an emergent conceptual and theoretical framework grounded in our multiple disciplinary pillars, the solution lies in reconciling these two perspectives: it involves considering the possibility of a gradual emergence of Darwinian properties in out-of-equilibrium systems subject to the fundamental constraints of thermodynamics.

We hypothesise that such a reconciliation could rest on autocatalysis and its ability to generate dynamics that are initially ecological but could gradually evolve into Darwinian dynamics. Our recent tools enable the exhaustive detection of autocatalytic cycles in large reaction networks and the analysis of their dynamics through simulations, while respecting the constraints imposed by thermodynamics. The first step of the project will aim at characterizing the parameters affecting the growth rate of these cycles, whether considered individually or collectively, using concepts and tools from theoretical ecology. In doing so, we will formally assess the inspiring proposition of a true ecology of autocatalytic cycles, based on a wide diversity of synergistic or antagonistic interactions between cycles. In the second part of the project, we will characterize the stationary states reached by these reaction networks when the production of each entity is balanced by its destruction. We will specifically seek to evaluate and understand the contribution of autocatalysis to multistability, i.e., the existence of multiple possible stationary states, each surrounded by its basin of attraction in concentration space. Indeed, multistability could serve as a physical analogue of heritable variance, paving the way for elementary Darwinian dynamics, the analysis of which will be at the heart of the third part of the project. By connecting multiple subsystems through diffusion, each occupying a compartment in a spatialized metasystem, we will test the hypothesis that such environmental structuring could give an advantage to the most prolific stationary states, rather than the most robust: not the most resistant to stochastic perturbations, but those most able to transmit their configuration to neighboring compartments. Thus, a stationary state characterized by a small basin of attraction could nonetheless invade the nascent population composed of these multiple connected compartments, for example, by producing, via autocatalysis, more matter than its neighbors. In this contrast between fitness and robustness, we glimpse the possibility of grasping the fundamental and general principles of the most elementary Darwinian dynamics. While purely theoretical, our approach is grounded in solid thermodynamic realism. In this sense, it constitutes a necessary and promising step toward the experimental investigation of these processes.