Is Sustained Macroevolutionary Progress Possible?

Life on Earth has advanced from prokaryotes to more highly organized forms including homo sapiens today. Thus life on Earth appears to have undergone "sustained macroevolutionary progress" ("SMP"). However, this did not occur in a bubble -- our planet is open to biological input from elsewhere. Life here may have been originally seeded by microorganisms from space. And lateral gene transfer that can be mediated by microorganisms plays a significant role in evolution. It is possible, therefore, that what appears to be SMP could actually be the development, over many generations, of pre-existing, highly evolved life from elsewhere.

Closed-System Tests Are Needed

If SMP is possible, it should be feasible to demonstrate it in the present. This has not yet been convincingly done. To ascertain whether SMP is possible, closed-system tests are needed.

Biology Is Better But Computers Are Faster

To test for SMP, closed-system biological experiments would of course be decisive. In fact, such experiments are under way. Pedigreed strains of bacteria have been bred for 25,000 generations now (PNAS 96:3807). Whereas mutation and recombination have been well observed, nothing in these experiments has given any promise of producing SMP.

Computers are analogous to biological cells in that they depend on lengthy instructions that are encoded in strings of symbols. Replication, transcription-translation (execution), mutation and recombination are operations that both systems share. If SMP could be demonstrated in a computer model that accurately mimics biology, the case for SMP in biology would be strengthened. And computers are a testing medium with one big advantage -- speed.

Several different computer models that mimic biological evolution have already been developed. Perhaps the best known is Tom Ray's Tierra. In runs of this model, smaller "organisms" tended to be the survivors. Furthermore, one organism occasionally "steals" a strand of code from another one, with a survival benefit. Parasitism such as this can contribute to evolutionary progress. However, without a supply of new instructions, parasitism alone will quickly reach an evolutionary limit. Tierra has not yet produced any wholly new instructions, nor evolved beyond its original parameters. Neither has any other model.

The challenge to design a computer model that could confirm SMP seems especially well-suited for NASA's Center for Computational Astrobiology.

A Confirming Example: Suppose . . .

Suppose a researcher has a copy of Richard Dawkins's Biomorphs program on a computer. Attached to the computer is a telephone, and the hardware -- but not the software -- for a modem. One day it happens that the Biomorphs, without prompting, have composed the software needed to operate the modem. Subsequently they escape via the telephone to other computers. Furthermore, other instances of this same phenomenon have occurred, and it has been duplicated in controlled experiments. Of course, this result would strongly confirm SMP in a computer model. One may object that the example is entirely farfetched because there are no incremental pathways by which to discover modem software from unrelated software. Granted. But does the same lack of incremental pathways prevent SMP in general?

Terms Need to Be Defined

Suppose a bacterium deploys genes it already possesses to metabolize a sugar different from the one on which it previously subsisted. Whereas adaptation has occurred, macroevolutionary progress has not. On the other hand, the first appearance of eukaryotes -- or multicellularity, cell specialization, oxygen metabolism, hard shells, feathers, bones, skin, teeth, lungs, kidneys, brains, or a long list of other organs and features -- each clearly represents macroevolutionary progress. If one such development is followed by others, the process appears to be sustained. Can these concepts be made precise enough to be verifiable on a scale observable within a human lifetime, either in biology or in a computer model?

New evolutionary features require new genetic instructions, in the same way that additional computer capabilities require new software. Could evolutionary progress be correlated with the size of the genome, or the size of the necessary genome, or the "least algorithmic complexity" of the genome needed to produce the phenotype? Such suggestions have not been well-received to date. Refining precise-enough terms to answer our question is a worthy goal in itself.

Is a Logical Proof Available?

Perhaps the question could be posed in a way that would make it susceptible to logical proof in the manner of Gödel or Turing. For example, "In any system based on encoded instructions, can the system itself, with no apparent limit, cumulatively compose and execute new instructions that are analogous to genetic instructions for macroevolutionary inventions?" Or, "can meaning come from nonmeaning?" A valid proof that SMP either is possible, or is not, would be monumentally important for biology and philosophy.

A Michelson-Morley Experiment?

In 1881, American physicist Albert Abraham Michelson (1852-1931) began a series of experiments to detect the presence of the luminiferous ether, the medium in which waves of light were supposed to be propagated. The tests were made increasingly precise and elaborate until the series culminated in 1887, when the definitive experiment was conducted with American chemist Edward William Morley (1838-1923). To widespread surprise, the ether could not be detected. This negative result helped open the way to the theory of relativity and a major scientific revolution.

Today, sustained macroevolutionary progress has not been demonstrated. Until it is, we cannot be sure that it is possible. Therefore, tests that could confirm the phenomenon are advocated. The issue is important because the first step in science is to ask the right question. The question has been, "How does life originate and evolve?" But we do not even know if it originates and evolves. Perhaps instead, life arrives and develops.

Darwinian Evolution Can Produce Adaptation

Evolution by expressing genes already in the genome:

Some adaptations require no genetic changes at all, but merely the expression of genes already in a species' genome. For example, the well-known English moth adapts its wing color to blend with soot-darkened walls and tree trunks by expressing a pigment gene already in its gene pool. Similarly, the African trypanosome has dozens of genes for different protein coats, and it deploys a new one as often as necessary.

Evolution by acquiring genes:

Other adaptations are accomplished by the acquisition of whole genes from elsewhere, called "lateral gene transfer." For example, bacteria often become resistant to drugs by acquiring "resistance plasmids" one or several genes long, directly or indirectly from other bacteria. Genes useful to bacteria are also installed by viruses in a process called transduction. Furthermore, viruses can move genes around in eukaryotes. Over 1% of the human genome is now recognized as endogenous retroviruses, and over 35% of our genome consists of transposable elements that may have been originally acquired by lateral gene transfer.

Evolution by point mutation:

During gene replication, errors occur at a low rate. In general the rate of errors per nucleotide replication is about 10-9 in prokaryotes and 10-12 in eukaryotes. But even this low rate can be useful. A single nucleotide substitutions can cause change quantitative changes such as shifting the coelacanth's visual sensitivity toward the blue end of the spectrum. Additional nucleotide substitutions can ultimately change even a majority of a gene's nucleotides without changing its function. This process has occurred in the highly conserved cytochrome-C gene, for example. The differences may serve to optimize each allele in a given species.

And a single nucleotide insertion or deletion can cause a frame-shift error that completely disables a gene. If the disabled gene is a control gene that normally promotes a cascade of other genes into action, the whole cascade will usually be disabled. If, however, the disabled gene is a suppressing control gene, the cascade would be enabled. In either case, the effect of this disabling mutation may be adaptive. In the latter case, if the newly enabled genes are activated for the first time in a species, it represents a step in macroevolutionary progress.

Gene duplication followed by mutation:

Whole genes can be duplicated, and once there are two copies, they may mutate differently. Gene duplication followed by a single point mutations in one allele appears to be the process whereby howler monkeys with dichromatic vision acquired trichromatic vision, for example.

Random changes in access codes:

Mutation and recombination in the genes for a viral protein coat may produce seemingly limitless variations that keep the virus a step ahead of the host's immune system. Here the immune system acts like a security expert (T-cell) seeking the password or access code (the antibody) to a password-protected subroutine (the virus). The virus survives by changing the access code frequently. Access codes are best when random, so random mutation and recombination can adequately produce them.

Can Darwinian Evolution Produce Sustained Macroevolutionary Progress?

For evolutionary progress new genes are required

The size of prokaryotic genomes ranges from about 800,000 to several million nucleotide pairs. As genes average about 1,000 nucleotides in length, bacteria have from, say, 800 to a few thousand genes. For example, a typical E. coli bacterium has about 4,700 genes.

At the other end of the evolutionary spectrum, the human genome has about six billion nucleotide pairs. However, eukaryotic cells are diploid, so half of these are merely "backup" copies. Furthermore, less than 5% of the human genome is in translatable gene sequences. Still, the number of genes in a human is estimated to be 100,000 or even 140,000.

Humans have organs and systems that prokaryotes do not. These features are made possible by genetic instructions that humans have and prokaryotes do not. Most human genes have no known analogs in the world of bacteria. Compared with any prokaryotic gene, they have nucleotide differences in dozens to hundreds of essential positions.

Ohno's method

Japanese-American pathologist Susumu Ohno (1928-2000) held that a new gene arises when an existing gene is duplicated and the new allele, while silent, undergoes mutations that give it a new function. But after one allows 10 or 12 point mutations at random positions, the number of possible gene sequences of average size (1,000 nucleotides) becomes larger than the total number of genes that could have been tested on Earth in 4 billion years. This process might work if almost any gene will do a "many worlds" theory of biology. But the process has not been demonstrated in any experiment.

Eigen's method

German physicist Manfred Eigen (b. 1927) suggests that new genes may arise by point mutations in existing genes that are not duplicated nor silent. In his analysis, a beneficial gene differing from an earlier gene by five nucleotide substitutions can be fixed in a species if the population is large enough, and if none of the intermediate genes is fatal. This method would require evolution to be ultragradual, which is not confirmed by the fossil record. Furthermore, it would require that there are gradual pathways, without any fatal intermediates, such that they lead from an original set of prokaryotic genes to, ultimately, all the genes employed by all higher life forms. This method of composing new instructions has not been demonstrated, in biology or computers.

New instructions required

The demonstrated methods of Darwinian evolution do not account for new genes that have lengthy new sequences with new instructional content. The methods proposed to account for them have not been theoretically developed, nor have they been demonstrated. The best example of an entirely new gene that arises from an existing gene comes from Antarctic cod. The gene for blood antifreeze appears to have arisen from a gene for trypsinogen, a pancreatic enzyme, by a series of duplications and recombinations. But the example is too rare to confirm SMP by Darwinian evolution.

Strong Panspermia:

Strong panspermia holds that in addition to life originally, the genes necessary for sustained macroevolutionary progress come from space. In this theory, what looks like sustained macroevolutionary progress on Earth is actually the development, over many generations, of pre-existing, highly evolved cosmic life.

Strong panspermia accords well with these phenomena that have troubled Darwinism:

  1. Life's rapid start on Earth
  2. Punctuated equilibrium
  3. Convergent evolution
  4. The ubiquity of certain master control genes
  5. The fact that many genes appear older, by sequence analysis, than they should be according to the fossil record

Strong panspermia is further supported by several lines of evidence:

  1. Organic compounds in space
  2. Fossils in meteorites from Mars
  3. Fossils in meteorites not from Mars
  4. The apparent immortality of some bacteria
  5. The newly recognized importance of lateral gene transfer in evolution

Perceived Difficulties for Strong Panspermia

The precept underlying strong panspermia is that sustained macroevolutionary progress (SMP) is not possible. It is commonly thought that the mere fact of evolution on Earth disproves this precept. But because the biosphere is biologically open, this is a naive rebuttal. Furthermore, the evolution of human phenomena such as architecture, science, or computer technology may be thought to disprove the precept. After all, computer programs cannot come from space! Yet this argument does not address the possibility that all our activity may represent the "emergent properties" of our genetic makeup.

The Origin of Life

The origin of life would be a prime example of sustained macroevolutionary progress, because the simplest living thing, a prokaryotic cell, is already quite complex. Therefore, any demonstration that life can originate from ordinary chemicals by natural means would answer our original question with a firm yes. Conversely, if SMP is not possible, then it follows that life cannot originate by natural means. This would mean that, without a miracle, life would have to come from the infinite past. This brings up the next problem . . .

The Big Bang

If the universe is a permanently closed system that began in a lifeless state a finite time ago, then sustained macroevolutionary progress, including the origin of life, must have subsequently happened in it. Most Darwinists adopt this premise and conclusion without question. Interestingly, creationists also endorse this premise and conclusion, but with a different motive. This unlikely coalition makes a strong lobby for this accepted version of the big bang theory.

But if sustained macroevolutionary progress is possible, science ought to be able demonstrate it in controlled, closed-system experiments. If it cannot be demonstrated, perhaps SMP is impossible. Without a miracle, this impossibility would indicate that there is a flaw in the premise. This reasoning reverses the usual roles and asks cosmology to accommodate a fundamental biological principle. In other words, if SMP is impossible, the universe can not be a permanently closed system that began in a lifeless state a finite time ago. We suspect that cosmology can make this accommodation without entirely dismantling the big bang theory.

Presented at the First Annual Astrobiology Science Conference
NASA Ames Research Center, Mountain View, California
3-5 April 2000

by Brig Klyce
Acorn Enterprises LLC