Monday, April 23, 2007

Sequence Similarity and Intelligent Design Creationism

Logan Gage posted a message on Evolution News& Views where he discusses the interpretation of sequences similarity [What Exactly Does Genetic Similarity Demonstrate?].
As Francis Collins, head of the project which mapped the human genome, has written of DNA sequence similarities, “This evidence alone does not, of course, prove a common ancestor” because an intelligent cause can reuse successful design principles. We know this because we are intelligent agents ourselves, and we do this all the time. We take instructions we have written for one thing and use them for another. The similarity is not the result of a blind mechanism but rather the result of our intelligent activity.
This is an old argument. It ignores the fact that sequence similarities match the phylogenies determined independently from comparative morphology and the fossil record. This is the "twin nested hierachies" evidence for evolution and it's powerful evidence indeed. Furthermore, it ignores the fact that the differences in sequences correspond closely to what we expect from evolution by random genetic drift.

In order to sustain the argument that an intelligent designer is responsible for this data you pretty much have to argue that the designer (whoever that might be) went out of his way to deceive us into thinking it's due to evolution. But that's not why I'm commenting on this article.
Some design proponents think the evidence for common ancestry is good (e.g., Michael Behe), while others—citing the fossil record, especially The Cambrian Explosion—do not. But neither group thinks that sequence similarity alone proves either common ancestry or the Darwinian mechanism, as so many science writers of our day seem eager to assume.
I congratulate Logan Gage for acknowledging that there are disagreements within the Intelligent Design Creationist camp. That's not something we see very often. However, I think he may be distorting Behe's position a little bit. Perhaps he means to place all the emphasis on "similarity alone" but that seems to be a quibble. Here's what Behe says in Darwin's Black Box on pages 175-177. Judge for yourself whether Behe thinks sequence analysis provides strong support for common descent.
When methods were developed in the 1950s to determine the sequences of proteins, it became possible to compare the sequence of one protein with another. A question that was immediately asked was whether analogous proteins in different species, like human hemoglobin and horse hemoglobin, had the same amino acid sequence. [The question was asked because it was a prediction of evolution-LAM]. The answer was intriguing: horse and human hemoglobins were very similar but not identical. Their amino acids were the same in 129 out of 146 positions in one of the protein chains of hemoglobin, but different in the remaining positions. When the sequences of the hemoglobins of monkey, chicken, frog, and others became available, their sequences could be compared with human hemoglobin and with each other. Monkey hemoglobin had 5 differences with that of humans; chickens had 26 differences; and frogs had 46 differences. These similarities were highly suggestive. Many researchers concluded that similar sequences strongly supported descent from a common ancestor.

For the most part it was shown that analogous proteins from species that were already thought to be closely related (like man and chimp, or duck and chicken) were pretty similar in sequence, and proteins from species thought to be distantly related (such as skunk and skunk cabbage) were not that similar. In fact, for some proteins one could correlate the amount of sequence similarity with the estimated time since various species were thought to have last shared a common ancestor and the correlation was quite good. Emile Zuckerkandl and Linus Pauling then proposed the molecular clock theory, which says that the correlation is caused by proteins accumulating mutations over time. The molecular clock has been vigorously debated since it was proposed, and many issues surrounding it are still contended. Overall, however, it remains a viable possibility....

The three general topics of papers published in JME [the Journal of Molecular Evolution]—the origin of life, mathematical models of evolution, and sequence analysis—have included many intricate, difficult, and erudite studies. Does such valuable and interesting work contradict this book's message? Not at all. To say that Darwinian evolution cannot explain everything in nature is not to say that evolution, random mutation, and natural selection do not occur, they have been observed (at least in cases of microevolution) many different times. Like the sequence analysts, I believe the evidence strongly supports common descent. But the root question remains unanswered: What has caused complex systems to form? No one has ever explained in detailed, scientific fashion how mutation and natural selection could build the complex, intricate structures discussed in this book.
The reason for bringing this up is to show that Behe accepts common descent and sequence similarity is strong evidence of common descent. It would be nice if the Intelligent Design Creationists acknowledged this and discussed the sum total of the evidence and not just the sound bite version of sequence similarity.

[I'd like to make it clear that I do not support everything Behe says. It's become clear to me recently that I need to add disclaimers such as this whenever I make a complicated point.]


  1. When you factor in the shared errors between species, the verdict becomes clear.

    Common descent is true, whether or not it happened purely by mutation, selection and drift.

  2. What creationists and their ilk are motivated to ignore or misunderstand are the processes by which sequences change (along with the evidence for such processes). Rates and processes (analog of calculus) are not easily grasped by those with weak math backgrounds. As long as arguments are restricted to states (not processes) the design argument retains plausability.

  3. "As long as arguments are restricted to states (not processes) the design argument retains plausability."

    Another method to recover enough of the collapsed description to see the implications (or better, test the predictions) is to look at the structure of nested hierarchies that the states embeds into. Sort of halfway picture of the process between a simplified state and a detailed rate model.