When most people, including scientists, even including evolutionary and molecular biologists, talk about evolution, they dwell on the critical aspects of variation and natural selection almost exclusively. These two processes are considered to be the heart and soul of the evolutionary process. But there is another process that must also exist for evolution to occur, and that is replication of the organism in the next generation. I am not talking about reproduction, which of course must occur. Replication is more than simply an organism reproducing itself. Replication is the means by which reproduction occurs with extreme accuracy, so that the characteristics of the original are inherited in the offspring. Replication is what makes inheritance possible. And the fact that replication is less than 100% accurate is what allows for the existence of variation, which is critical for evolution.
What we tend to overlook is that replication, while it cannot be perfect (so as to allow for variation), must be very, very good. If it weren’t, then whatever selective advantage an organism might have gained by a change in some allele would not be transmitted to the next generation, and no evolution would occur. If a bird had developed better vision than its siblings, that bird would have a great selective advantage during its own lifetime. But if that characteristic were not inherited by its offspring, evolution of better eyesight in that species would never happen.
Modern life replicates its phenotype with at least 99.99999% fidelity. This leaves enough room for naturally occurring errors (mutations) to produce the variation needed for Darwin’s theory to work. But what about the lower limit of replication fidelity? How good must replication be in order to avoid “error catastrophe”—which means, in this context, a level of error such that no selective advantage is possible?
The threshold for a mutation rate that would cause an error catastrophe has been determined theoretically and confirmed by experiment to be simply equal to the inverse of the size of the genome. Thus, if an organism has a genome of 10,000 bases, like some bacteria and viruses, a mutation rate greater than 0.0001 or 0.01% would lead to a loss of any selective advantage for the fittest organisms, and thus it would not allow for evolution to work. This seems like a very low mutation rate, and it is, but of course in large multicelled organisms with genomes in the billions of bases, the error rate is correspondingly lower. Since replication fidelity is equal to 1 minus the mutation rate, the minimal level of replication fidelity is 0.9999 for single-celled organisms, and 0.9999999 for animals and plants.
Are such high values for replication fidelity in early life reasonable to expect? Not when we consider that for even the most primitive modern organisms, replication, transcription and translation involves a host of error correction enzymatic processes, all of which had to evolve—but how could they if the prevailing error rate was too high to allow for evolution?
This seems to leave a large gaping hole in our attempt to understand the origin of life, particularly the origin of evolution. The best solution to this mystery is to posit some other type of evolutionary process whereby early primitive cells could replicate themselves (meaning their entire phenotype) with great accuracy that did not involve the extremely complex advanced mechanisms of DNA replication and DNA-directed protein synthesis. The RNA world (generally assumed to predate DNA world) doesn’t look much better. Even if we assume that the RNA-world genome is a set of RNA ribozymes, and that the smallest such self-replicating molecule might be as small as 50 bases, we still need to have a 98% accuracy in replication of RNA, which is far less than the 80% fidelity (at best) observed in lab experiments. We have no idea what such an alternative evolutionary mechanism not working with replicating nucleic acid polymers might consist of.
But we can still address the fundamental question of replication fidelity evolution even if we have no idea how that evolution could have occurred. I recently completed a study of a simulation model for studying replication fidelity in early life that makes no assumptions about replication mechanisms.
The results (which I have recently submitted in a paper for publication) are interesting. To summarize, it seems very clear that regardless of what the unknown evolutionary mechanism might be, a smoothly continuous evolutionary path to high replication fidelity is impossible. At some point during the evolution of protolife to modern life, there had to be one or more major jumps (saltational events) in the degree of replication fidelity. We have a long way to go before we can get close to any idea of how life and evolution might have gotten started.