Four Billion Years
An Essay on the Evolution of Genes and Organisms

by William F. Loomis

This is a review of Four Billion Years: An Essay on the Evolution of Genes and Organisms, by William F. Loomis (Sinauer Assoc., Inc., 1988, ISBN 0-87893-476-6). It is a look at organic evolution from the perspective of a biochemist.

Like any book of evolutionary history, this one records the step-by-step appearance of new forms, slowly building up to the repetoire we are familiar with today. Unlike most, it concerns itself mostly with forms of genes and proteins, not anatomies. The book is full of reconstructed genealogies of proteins. For instance, we learn that the globins are an ancient family of proteins antedating the appearance of the plant and animal kingdoms. Or so we must infer from the presence of hemoglobin in both animals and plants such as legumes.

One interesting feature of the book is that it is closer in proportion to the actual timeline of life. Most books on evolution say, in effect, "Life arose four billion years ago. Half a billion years ago, we find the first significant fossils..." and go on from there. Loomis doesn't even get around to eukaryotes (nucleated cells) until page 157, or to multicellular life until page 185. (There are 264 pages in the main text.)

The price paid for this original approach is that he has very little direct evidence of the period of life he spends the most time on (the Precambrian). He must rely on comparative genetics, to infer the nature of ancestral genes from their surviving descendants, assisted with, as he freely admits, a lot of conjecture.

For it is genes and proteins he is mainly interested in. And he writes for people who are interested in them and conversant with biochemistry. He does not, for instance, stop to explain what cDNA is, or mRNA, or operons, or pyruvate. (Personally, I was able to figure out cDNA from context, sorta, didn't really care much about pyruvate, and happened to remember mRNA and operons.)

Loomis's hope is that genetic genealogy such as he displays will eventually be able to group the estimated ten million genes in circulation into a few large family trees.

For me, the most interesting parts of the book were the beginning and the end. At the beginning, he produces the latest speculations on the origins of life. At the end, in an appendix, he produces an imaginary survey of what we might find on ten randomly selected Earth-like planets.

Loomis presents the following sequence for the origins of life:

In the beginning is RNA, formed from various chemicals plausibly present on the young Earth. It has been recently discovered that, besides carrying genetic data, RNA can sometimes act as a catalyst on other RNA, promoting cleaving and replication (though not polymerization, at least not yet observed), and can "edit" itself.
Loomis then pictures short lengths of this RNA as picking up amino acids and binding to other lengths of RNA, the amino acids then binding into proteins. (This is, in fact, protein synthesis as it occurs in the ribosomes of living cells, but without the assisting ribosomal RNAs and proteins that make up the rest of the organelle.) If the protein thus produced acts as an enzyme assisting the protein sythesis in any way, you have a chemical system termed an "autocatalytic cycle."
Autocatalytic cycles may then make products that assist, not themselves, but each other. Such systems are dubbed "hypercycles."
These cycles could be easily disrupted by simple dilution, but could be concentrated in microscopic bubbles of lipids. These lipids form in the presence of metal ions, from carbon monoxide and hydrogen gas. The bubbles are about the size of bacteria, essentially little soap bubbles, filled with water and suspended in water. The lipid membranes tend to pick up proteins and nucleic acids.
A lipid bubble that acquires a robust set of hypercycles and encounters ("feeds on") a sufficient quantity of amino acids could then accidentally split and still have sufficient quantities of all the RNAs and proteins to continue the hypercycle in both daughter drops. This is still too haphazard for Loomis to call it metabolism and reproduction, but it is clearly on the way.

Loomis then goes on to speculate on how this semi-animate metabolic cycle could pick up ATP as an energy transfer agent, develop a home-made cell wall instead of trusting to chance encounters with soap, transfer genetic information to DNA as a kind of molecular archive, and so forth through mitosis, photosynthesis, respiration, and nucleated cells.

He rightly points out that the development of multi-cellular organisms and fiddling details like bones and brains are really fairly minor design changes compared with all that had gone before, in the world of the bacteria and other prokaryotes. So no wonder all that kind of thing has merely taken the last 10% or so of Earth's history.

At the other end of the book, we have the survey of ten planets. For each life-bearing planet, he details the metabolic differences, such as whether the sugars or amino acids are right-handed or left, or in which direction the local life reads its genetic code. Of the ten, only three have appreciable quantities of free oxygen. These are, not coincidentally, the ones with macroscopic, multicellular life (in one case, sapient). Of the remaining seven, one used to have life and six harbor nothing but pond scum. Of course, by now the reader should be quite a connoisseur of pond scum and recognize it for the achievement it is.

Plausibility

With just a little forcing, I could present Loomis's speculations on biogenesis as an unintentional but eloquent case for creationism.

Without forcing, I feel it shows how little is known, and how much a preference for a purely physical origin of life depends either on a prior metaphysical preference or on sticking to the rules of the game of science and not allowing in things that lie outside the proper sphere of science.

I am particularly nonplussed by the picture of ribosomal protein synthesis going on without benefit of ribosomes and just happening to produce proteins that assist that synthesis (step two in the outline in the previous section).

First of all, the protein that gets synthesized is a random one. What are the odds that it will be helpful? or harmful (e.g. something that digests RNA!)? or irrelevant? Can we even compute those odds?

So far, every argument I have ever seen for or against natural biogenesis has failed to compute these odds, or even show that they might be computable.

Second, it's rather doubtful that the protein is really coded in any of the RNA. For that to happen, there would have to be specific affinities between the different forms of proto-tRNAs and different amino acids. Loomis's own notes remark that

Gamow (1954), Woese (1966), and others have tried to see how nucleic acids might code for specific amino acids on chemical grounds. No convincing interactions that would be sufficient to account for the nature of the code were found. Jukes (1983) has considered a primitive code specifying on 16 amino acids. The interactions suggested in this essay are purely speculative and are presented only to raise the possibility that under some conditions limited specificity may have played a role. The function of simplified RNA molecules are described by Kinjo et al (1986).

(In living cells, there is a set of enzymes for matching particular tRNA molecules to particular amino acids.)

Third, assuming a helpful protein to get synthesized, it will only be helpful for part of the process, in all likelihood. That is, it may help bind amino acids to tRNAs, or it may help polymerize RNAs, or polymerize amino acids, or help RNA to replicate, etc., but it will only do one of those jobs. That's why the theorists introduced hypercycles – one autocatalytic cycle for each of these jobs, assisting each other and forming the hypercycle.

Unless the autocatalytic cycles evolve from one another, you only get to hypercycles by having several autocatalytic cycles appear in close proximity to one another

If you set yourself the question, "How could life arise on Earth by natural * causes?" then you will, of course, come up with an answer like the one I have been criticizing. And that is, perhaps, the only kind of question science could ask about the origin of life.

(* Unnatural causes can be divided into artificial and supernatural. Supernatural is the classic creationist position. For a look at the artificial option, there's panspermia.)

Comment from Wally Neilsen-Steinhardt:

The detailed mechanism in [the review] does seem to involve a few very small probabilities. I would therefore entertain the hypothesis that some other mechanism(s) may actually have occurred. This hypothesis cannot currently be proved, but it can be scientifically investigated. There are some investigations going on right now which suggest that other mechanisms would have higher probabilities.

Those who want to refute evolution or the origin of life by natural means must show that no natural mechanism could produce evolution or life with a probability significantly different from zero.

Reply:

If the probability turned out to be small, wouldn't it be necessary to go out and perform an actual survey of the incidence of life in the universe?

If we could compute the a priori odds and they were high or reasonable (say they were .99 or at least .10 or .05), we could conclude that we were perfectly ordinary or only modestly lucky and go on. But if the a priori odds were reliably known to be ten-to-the-negative-horrid, we would still not know if our existence was in any way unnatural unless we batted around the universe and found that there were far too many life-bearing worlds for the odds.

Now suppose the a priori odds were reliably known to be large and we went exploring and found too few life-bearing planets.... (Evidence for what Jim Burrows, who is a fan of Henri Bergson, dubbed the "elan fatal.")

Options

I've said that the supernatural creation hypothesis is probably not a scientifically good one for the origin of life, because it seems hard to test empirically. But suppose you aren't limiting yourself to the science game. How does the creation hypothesis fit into general philosophical considerations?

There can be no very neat answer, because general philosophy, unlike science, is characterized by a gross lack of concensus on lots of issues. But here are some of my opening thoughts on the subject:

First, there is not one creation hypothesis, but many. One could devise one for every creation myth/theory/account/speculation ever coined, but the interesting ones, of course, are the ones attached to the monotheistic religions. We can break the creation hypothesis into its several variations by asked the additional question, "How did God create life?" Some (only some) of the answers are:

Hard-Core Creationist: He created the Earth and all the species on it much as they are now, 6000 years ago. The physical evidence of this is plainly visible, but the atheistic scientific establishment conspires to brainwash itself and the public into ignoring it.

Rigged-Demo Theorist: He created the Earth and life 6000 years ago, but gave them the appearance of having been developed over many millions of years.

Soft-Core Creationist: He created the Earth and life over as much time as you like, using some combination of providential and miraculous guidance.

Stacked-Deck Theorist: The naturalistic account of the origin of the Earth and life is correct as far as it goes, but it is God who designed the universe to generate Earth and its life through its laws and boundary conditions.

Atheist: He didn't.

It is easy to argue with the hard-core position, but hard to get anywhere, or to make much philosophical use of it except as a springboard to such general questions of science philosophy such as "what is evidence and how do you weigh it?" Any other fringe-science or pseudo-science position would serve as well.

The rigged-demo theory is clearly designed to isolate the doctrine of creation from any empirical considerations. (Historically, it is known as the "omphalos" theory, from the Greek for navel, referring to the question of whether Adam was created with a navel.)

The atheist position has variations within it like the others, but can often be equally proof against empiricism. It often takes science and inflates it into the whole of philosophy. Thus, if God is not a fruitful scientific hypothesis, He is not a legitimate philosophical hypothesis, and one is virutally not allowed to consider any evidence as pointing in His direction.

The soft-core and stacked-deck positions are, I think, amenable to influence by empirical evidence. So might some kind of soft-pedaled agnostic position, or a panspermist position similar to Hoyle's.

For instance, if you are trying to decide betwen soft-core and stacked-deck, and you learn that the a priori probability of life appearing spontaneously on Earth is rather high, this should push you toward the stacked-deck position, since this makes miraculous intervention seem superfluous. (Does God need to work a miracle to make the sun come up in the morning?)

But the situation isn't symmetric. If you learn the odds are low, this might be because the soft-core position is true, or it might be because the stacked-deck position is true and:

a) God stacked the boundary conditions more than the laws, so there are more life-bearing worlds than the odds would countenance.
or
b) there aren't many life-bearing worlds, but we're one of them.

It might, however, be hard to distinguish between stacked-deck-a and soft-core.

Of course, we don't know the odds, and we don't know the actual frequency of life-bearing worlds. But they are things we might conceivably come to know.