What is life? Ostensibly, this question has as many answers as the number of philosophers who have attempted to define life. Yet, they all agree in one respect that all “living” things are a self-replicator, i.e., if the conditions are right, they can create a nearly identical copy of themselves. That’s why any attempt at understanding the origin of life tries to create a self-replicator in vacuo. There are two schools of thought regarding the creation of an artificial self-replicators that approach the problem from opposite end of the spectrum. Despite their differences in approach, both schools agree that creation of self-replicators requires (1) at least one autocatalytic cycle of the reactants that consumes chemical fuel to create the copy in a repeated fashion and (2) a steady source of the chemical fuel that sustain the autocatalytic cycle; if the autocatalytic cycle is fully functional it leads to exponential growth in the number of the self-replicators. Agreeing on this point, the two schools diverge from each other. The first school engineers an autocatalytic cycle by hand and attempts to tune the chemical parameters to sustain the autocatalytic cycles. The second group designs a concoction of chemicals, where autocatalytic cycles emerge spontaneously.
Unfortunately, neither group has been able to create a self-replicator that grows exponentially. This lack of success perhaps has its root in trying to create self-replicators from fundamentally unsustainable autocatalytic cycles. Indeed, a theoretical calculation shows that even creating a three-step cycle, e.g., A→B→C→A, requires the cycle to be completely isolated from any parasitic reactions that consume A, B, or C [1]. However, self-replicators are quintessential. It is through self-replication that the whole living world sustains itself. Perhaps, the most efficient mode of self-replication is not through a single autocatalytic cycle. Indeed, a recent experiment by Vaidya et.al. [2] shows that self-replication tends to happen more efficiently through the cooperation of multiple reaction cycles that by themselves need not be autocatalytic, but when taken together they allow for autocatalytic growth.
This is where our work begins. We were trying to create a chemical library where self-replication emerges spontaneously. We were trying to tune some high-level parameters, such as the interaction strength between the molecules that constitute the reactions in our chemical library. We could enumerate all isolated autocatalytic cycles with 2 or 3 steps. Because our chemical system contained so many possible parasitic reactions that we could not imagine getting self-replication out of 4 or higher step autocatalytic cycles. We eventually identified a set of parameter values where we could see exponential growth. However, to our great surprise, we could not find a single isolated autocatalytic cycle that could sustain exponential growth.
Something was amiss. This is when we came across the paper by Vaidya et.al. [2], and we guessed that perhaps in our system self-replication is happening through cooperation of multiple cycles. Further analysis showed that it was indeed the case. In fact, these cooperating cycles could sustain themselves even when the parasitic reactions allow only 0.01% of the molecules to react in the reactions that constitute the self-replication reactions! We found that self-replication through multiple cooperating cycles is a norm in most chemical systems. Our suggestions to future designers of self-replicators is that focus on engineering a chemical system where self-replication occurs through cooperating cycles. You will soon land an artificial self-replicator.
References:
[1] Szathmáry, Eörs. “The origin of replicators and reproducers.” Philosophical Transactions of the Royal Society B: Biological Sciences 361.1474 (2006): 1761-1776.
[2] Vaidya, Nilesh, et al. “Spontaneous network formation among cooperative RNA replicators.” Nature 491.7422 (2012): 72.