Reviewer 4: Itai Yanai (Harvard University)
In this work, Eugene Koonin estimates the probability of arriving at a system capable of undergoing Darwinian evolution and comes to a cosmologically small number. With such an improbable event at hand, Koonin turns to a cosmological perspective in order to grasp its feasibility. He cites recent work in cosmology that highlights the vastness of the universe, where any series of events is necessarily played out an infinite number of times. This so-called "many-worlds in one" model essentially reconceives any chance event as a necessary one, where its (absolute) abundance is proportional to its chance of occurring.
The context of this article is framed by the current lack of a complete and plausible scenario for the origin of life. Koonin specifically addresses the front-runner model, that of the RNA-world, where self-replicating RNA molecules precede a translation system. He notes that in addition to the difficulties involved in achieving such a system is the paradox of attaining a translation system through Darwinian selection. That this is indeed a bona-fide paradox is appreciated by the fact that, without a shortage effort, a plausible scenario for translation evolution has not been proposed to date. There have been other models for the origin of life, including the ground-breaking Lipid-world model advanced by Segrè, Lancet and colleagues (reviewed in EMBO Reports (2000), 1(3), 217–222), but despite much ingenuity and effort, it is fair to say that all origin of life models suffer from astoundingly low probabilities of actually occurring.
Koonin's main contributions in this manuscript are two-fold: 1. a description of a minimal "breakthrough system" capable of priming Darwinian evolution" and most importantly 2. relating the issue of overcoming probability barriers by defaulting to the anthropic principle, which is supported by advances in cosmology. Together these provide for a model of the origin of life where the "breakthrough system" appears by chance and is sufficient to prime Darwinian evolution. Koonin distinguishes between a strong and weak form of this model. In the strong form, the entire "breakthrough system" occurs entirely by chance. While in the weak form, a less complex system (and thus more cosmologically common) is found that is able to achieve Darwinian evolution. Should such a less complex system be discovered, the breakthrough system as Koonin describes it will have been falsified; however, even this less complex system is likely to be vanishingly rare and consequently also requires the anthropic principle to account for its occurrence on Earth. Of course, the model would also be falsified if the "many-worlds in one" model is itself falsified.
Overall, this is a bold manuscript that promises to deeply influence the stream of thought on the origin of life. To my knowledge the present model represents the first one to account for the origin of life by explicitly invoking the anthropic principle. Whereas the sufficiency of time has been questioned for evolving life, invoking the anthropic principle allows for an elegant – albeit science-fiction-like – way out of this chicken and egg problem.
From this perspective, future advances in the field of the origin of life may better estimate the ubiquity of life in the universe by attempting to break down the "breakthrough system" into less complex parts. In the very extreme, future work may show that starting from just a simple assembly of molecules, non-anthropic principles can account for each step along the rise to the threshold of Darwinian evolution. Based upon the new perspective afforded to us by Koonin this now appears unlikely.
Author's response: I agree with most of the statements in this constructive comment. Once again, however, I should note that, the way I see this situation, it is impossible to shun anthropic reasoning completely, whatever the advances of future work. As Yanai puts it, "in the very extreme", one could dream of non-anthropically explaining the entire sequence of evolutionary steps from monomers to a RNA-protein world. However, in the preceding history, an anthropic component inevitably will remain.
Perhaps, to complete the discussion, a final comment on the anthropic principle/selection/reasoning is due here. In all four reviews of this work, regardless of the other opinions of the reviewers, there is a strong emphasis on the anthropic principle that, in my view, is somewhat misplaced. Surely, the anthropic principle is important. However, I believe that it is secondary to the actual model of the uni(multi)verse. Indeed, the infinite repetition of all permissible histories in the MWO cosmology makes anthropic selection a straightforward epiphenomenon of the model (see text and Table 1). I should add that I also find it to be more satisfying philosophically that the model is put ahead of a "principle". Should the model be falsified, the status of the principle will become uncertain, and of course, the entire concept developed here, if not refuted in its entirety, will require a drastic revision (as rightly emphasized by Yanai).