Abstract
We determine conditions under which a random biochemical system is likely to contain a subsystem that is both autocatalytic and able to survive on some ambient 'food' source. Such systems have previously been investigated for their relevance to origin-of-life models. In this paper we extend earlier work, by finding precisely the order of catalysation required for the emergence of such self-sustaining autocatalytic networks. This answers questions raised in earlier papers, yet also allows for a more general class of models. We also show that a recently-described polynomial time algorithm for determining whether a catalytic reaction system contains an autocatalytic, self-sustaining subsystem is unlikely to adapt to allow inhibitory catalysation - in this case we show that the associated decision problem is NP-complete.
The Origin of Metabolic Engines
A central mystery in prebiotic chemistry is how a self-sustaining metabolic engine could emerge before the advent of complex genetic machinery like DNA or RNA. Stuart Kauffman’s "metabolism-first" hypothesis suggests that life began as a self-sustaining autocatalytic set—a collection of molecules where every member is produced by a reaction catalyzed by another member of the set, all derived from a basic "food" source.
While conceptually appealing, this theory required rigorous mathematical grounding to determine the likelihood of such networks forming spontaneously in a random molecular "soup." Mossel and Steel address this by identifying the precise thresholds of catalysis required for these networks to emerge and analyzing the computational complexity of identifying them in more realistic, inhibitory environments.
Defining Self-Sustaining Networks
The authors formalize these systems as Catalytic Reaction Systems (CRS), distinguishing between two primary types of self-sustaining sets:
- RAF (Reflexively Autocatalytic and F-generated): A set of reactions where every reaction is catalyzed by at least one molecule within the set, and every molecule can eventually be synthesized from the initial food source ($F$).
- CAF (Constructively Autocatalytic and F-generated): A more restrictive subset where molecules must be built up in a strict sequential order, ensuring each step is catalyzed by a molecule already present in the system.
Probabilistic Thresholds for Emergence
Using probabilistic combinatorics on random sequence-based models—where molecules are represented as strings of length $n$—the paper identifies a stark divide in the requirements for these two types of networks:
- RAF Threshold: For a RAF to emerge with high probability as $n \to \infty$, the average number of reactions catalyzed by a single molecule ($\mu(n)$) only needs to grow linearly with the molecule's length ($n$). This suggests that "looser" autocatalytic sets are surprisingly likely to form by chance.
- CAF Threshold: In contrast, the emergence of a CAF requires catalytic activity to scale exponentially with sequence length, matching the growth of the total number of possible reactions. This indicates that highly structured, sequential metabolic pathways are significantly harder to form spontaneously.
Complexity and Inhibition
In "monotone" systems where molecules only facilitate reactions, identifying a RAF is computationally efficient (solvable in polynomial time). However, real biochemical systems often involve inhibition, where certain molecules prevent specific reactions.
The authors prove that once inhibitory catalysis is introduced, the problem of determining whether a system contains a RAF becomes NP-complete. This result, derived via a reduction from the 3-SAT problem, suggests that finding self-sustaining networks in complex, realistic chemical environments is a fundamentally "hard" problem, potentially limiting the search space for both natural evolution and laboratory simulations.
Discussion and Future Directions
This work provides a robust mathematical framework for prebiotic chemistry, showing that the "metabolism-first" scenario is theoretically plausible under relatively modest catalytic requirements. However, the model focuses on network topology rather than reaction dynamics. Future research must bridge this gap by incorporating molecular concentrations and reaction rates. Additionally, the "plague of side reactions"—which can be modeled as a form of inhibition—remains a critical hurdle, as the NP-completeness result suggests that complex inhibitory networks are difficult to analyze and perhaps even more difficult to spontaneously assemble.
Summary updated