From earliest recorded history mankind has wondered how life on earth first arose. The current diversity of life on earth is spectacularly well-explained by Darwinian (or Dawkinsian) evolution, the process of replication with random variation plus natural selection. Things that are better at making copies of themselves make more copies. What makes something better at reproducing in one environment almost always makes it worse in a different environment. Lungs are a big win if you live on land, gills generally work better if you live under water. Earth has a large variety of environments, and so a variety of life has evolved to exploit them.
But this leaves a crucial question unanswered: how did this process get started? Even the simplest living thing today is far too complicated to have arisen by pure chance. There has been a lot of speculation and plausible hypotheses, but no actual answer. The belief that life began as a purely naturalistic process has always required a kernel of faith.
Until now. In the last year and a half there have been two papers published that have removed the last vestiges of reasonable doubt. The first one was a computer model, and the second, published just three months ago, is an actual laboratory experiment. (Here is a more accessible description.)
But before I describe these papers I want to show you a little back-of-the-envelope calculation that illustrates why they are so significant. To kick start evolution we need to somehow make a replicator, a thing capable of making copies of itself. In order to assess the likelihood that a replicator can arise by purely naturalistic processes we need to know two things. First, what is the most complicated thing that could plausibly arise by pure chance, without life? And second, what is the simplest plausible replicator? If there is a big gap between these two, then we have a big Problem, a big gap in our explanation.
A likely candidate for a minimal replicator is an RNA molecule because RNA is a biological multi-tasker: it can both carry genetic information and catalyze chemical reactions of the sort that happen in living things. RNA, like its close chemical cousin DNA, is a polymer, a molecule that consists of a chain of small building blocks called bases. Both RNA and DNA have four different bases. Three of these are the same: adenine, cytosine, and guanine. The fourth base in DNA is thymine while in RNA it is uracil. These are commonly abbreviated ATCG and AUCG, but these details don't really matter. What matters is that in each case there are four different bases, and in both cases these bases are arranged in a linear sequence. This makes it really easy to compute how many possible DNA or RNA molecules there are with a given length: it's just 4 raised to the power of the length of the chain. (Strictly speaking you have to divide this number by two because if you take a sequence and reverse it you end up with the same molecule, but that turns out not to matter.)
It has been experimentally demonstrated that the bases that form RNA (and DNA and proteins) form spontaneously in conditions likely to have existed on earth in its early days. It has also been experimentally demonstrated that these bases spontaneously link together to form chains. What had not been experimentally demonstrated until now was that these spontaneously generated RNA chains could form replicators. In fact, there seemed to be a pretty big gap between the complexity of the chains that had been formed in labs and what would be needed to self-replicate, but this was hard to assess because we didn't actually know how short a replicator could be.
It is pretty straightforward to predict what this value should be. We start by estimating how much material we could have to work with. Earth's current biomass, i.e. the total mass of all the organic compounds on earth is about 500 GTC (gigatons of carbon). Note that this is only a tiny fraction of the total carbon on earth. That figure is 1.85 billion GTC. Only about one in a million carbon atoms on earth are part of an organic molecule. So it is possible that the biomass of the early earth was much higher, but that will ultimately turn out not to matter.
The numbers we are about to deal with are going to get very big so it will be convenient to swtich to scientific notation. Unfortunately, the Blogger platform doesn't make it easy to create superscripts, so I am going to use the conventional 10^X notation to denote 10 raised to the power of X. 500 GTC is 500 x 10^9 = 5x10^11 tons = 5x10^14 kilograms of carbon. Let's be conservative and round this down to just 10^14 kg. To get the number of carbon atoms we multiply by Avogadro's number 6x10^23, and divide by 12 (because the atomic weight of carbon is 12 —six protons and six neutrons). Since we are just doing a very rough estimate here, we can safely ignore everything but the exponents and arrive at a final figure of (very roughly) 10^45 carbon atoms. The RNA/DNA bases all have less than six carbon atoms, so this is enough to make 10^44 RNA/DNA bases. Of course, not all organic molecules are RNA/DNA bases, so let's round this down to 10^40. That's dividing by ten thousand, which seems pretty conservative.
The other thing we need to take into account is how much time we have to find a replicator. How fast do these chemical reactions takes place? How long does it take to stick a new base onto an RNA chain, or take one away? We can get a rough estimate by looking at how long it takes for a living organism to reproduce. The well-known bacteria E. coli takes about 40 minutes to reproduce, and it has 4.7 million bases in its genome. That's about 1000 bases per second, but this is likely a serious overestimate for prebiotic earth. Life has had billions of years to optimize its reproductive chemistry, so let's be conservative and assume that it takes a full second to build a new RNA molecule in a prebiotic earth. There are 60x60x24x365 = 7x10^7 seconds in a year. Again, let's be conservative and round this down to 10^7. But then we need to multiply this by the amount of time we have to produce a replicator. Earth is four billion years old, so if we can do it in (say) a million years that is the blink of an eye on that time scale. So we have 10^40 bases, and 10^7 years which gives us time to do 10^14 different experiments. Note that this is not to say that we can only try 10^14 different combinations. All of those 10^40 bases are floating around in the primordial soup and mixing and matching and forming different sequences at the same time. So every second we can try a huge number of combinations. How many? That is a little tricky to compute because we don't know how long a sequence we actually need. Again, let's be conservative and look at the smallest currently existing natural replicators to guide us. These are called viroids, and they have a few hundred bases. Let's round this up to 1000. So every second we can potentially try 10^37 different sequences. Multiply that by 10^14 seconds and we can roll the abiogenetic dice a total of 10^50 times in a million years.
Is that enough? If the minimal replicator that we're looking for is roughly the same size as a modern viroid, i.e. a few hundred bases, then no, it's not enough. Not even close. And so for a very long time the abiogenesis hypothesis relied to a certain extent on an article of faith: a replicator that is much shorter than anything that exists on earth today is possible. Another way of looking at it is that this hypothesis made a falsifiable prediction that a much smaller replicator is possible.
For a long time there have been some good theoretical reasons for believing that shorter replicators are possible, but no actual experimental proof. These theoretical reasons have to do with information theory and the theory of computation and a theoretical construct called a Quine. That is a deep thicket of weeds that I want to avoid here, though it is all rather fascinating if you feel like diving in. The bottom line is that under some not-entirely-unreasonable assumptions you can demonstrate mathematically that self-replicating systems are possible with as little as 132 bits of information, which is the equivalent of 66 base pairs. That is easily in range of what can be achieved with 10^50 trials. The math goes like this: suppose you have an extremely unlikely event with odds 1 in N where N is a very large number. If you do exactly N trials then the odds that this event will occur is about 2 in 3 (the exact value is 1-1/e, about 63%). After that the odds rise dramatically. If you do 2N trials then the odds of the event happening are 87%. If you do 10N trials the odds rise to over 99%.
In other words: if you do 10^50 trials, and there exists a replicator whose odds of arising by chance are better than 1 in 10^49, then you are practically guaranteed to find it. Those are the odds for a biological replicator with about 80 base pairs (because 10^49 is approximately 4^80).
And now we have an actual experimental demonstration of an RNA replicator with 45 bases (it is called QT45). So even if we are off in our estimate of 10^50 trials by many, many orders of magnitude (and remember that we arrived at that number by making some very conservative assumptions) it is still a virtual certainty that a replicator will arise spontaneously almost immediately (on cosmic and geological time scales) on a planet with liquid water and a biomass the size of earth's.
It is important to be clear about what this experiment actually shows. It does not show that this is how life actually began. It does not show that QT45 is the original replicator. All this experiment shows is that small biological replicators are possible. But that is enough. If they are possible, and they are small enough (and QT45 passes the necessary threshold by a huge margin), then the spontaneous generation of replicators is inevitable. Abiogenesis no longer requires any leaps of faith. The details of how it actually happened are still TBD and probably always will be. But the mere fact that a naturalistic explanation has now been demonstrated to be possible beyond any reasonable doubt completely destroys intelligent design. As long as there was room for doubt that purely naturalistic abiogenesis was possible, there was room for a reasonable belief that some kind of intelligent designer was necessary. But that argument has now been blown out of the water. We no longer need to guess how small a replicator can be, nor do we need to guess how likely it was for one to arise by chance. Now we know.
Intelligent design advocates will object to this by pointing out that this replicator did not arise in nature but was created in a lab, which was created by intelligent humans. But this completely misses the point. What matters here is not how this replicator was created, but the fact that it was possible to create it at all. This replicator is almost certainly not the one that originally sparked life here on earth. It is almost certainly not the smallest possible replicator. It is almost certainly not the most effective replicator of its size. It is actually not a particularly good replicator. But it is also almost certainly not the last replicator of this size that we are going to find. The fact that one replicator this small exists means that it is virtually certain that there are others, and that some of them will be smaller, and some of them will be better. And yes, these things can be created simply by tossing the parts into bin and shaking them up — as long as your bin is the size of a planet and you shake for a few million years, though for obvious reasons that is not an experiment we are likely to be able to replicate.
The other paper, published last year, which demonstrated all of these things happening in a computer model is the icing on the cake. The results here are not directly comparable to a biological system. There are good reasons to believe that the computer model captures the dynamics of a biological system, but that is a very deep rabbit hole. But the main takeaway is that the replicators which arose in the computer model are of comparable complexity to the QT45 replicator.
This is the last nail in the coffin of intelligent design theory. Before these results, intelligent design could only be criticized as an argument from ignorance: just because we don't know the details of the process that produced the first replicator doesn't mean that it was not a naturalistic process. All it means is that we have not yet worked out the details. But now we have. The last gap in our understanding of the naturalistic origins of life on earth has now been definitively closed.