Wednesday, February 9, 2011

An illusion of an illusion of common descent

Over at BioLogos, Peter Borger has jumped in and posted a series of comments on Venema's latest critique of RTB. One of the responses to Borger was, "I must admit that I am confused by some of Peter Borger’s comments." Trust me; you're not the only one. Since Monday, I've gotten several emails about Borger's papers, and I think it's about time I commented publicly.

First, I should note that Borger has published a follow-up paper with Royal Truman on the GULO pseudogene, called "Why the shared mutations in the Hominidae exon X GULO pseudogene are not evidence for common descent." You can find the PDF from Journal of Creation right here.

I'm actually not going to comment on that paper. I'm far more interested in the paper Borger published in JC last year, "An illusion of common descent." There's no PDF online, but here's the reference (in case you have a JC subscription):

Borger. 2010. An illusion of common descent. Journal of Creation 24(2):122-127.

I mentioned the paper last year when the it came out, and then I just dropped it. Mostly because of my mom's advice about what to do when you can't say something nice. But now people are bugging me about this, so fine. If you want to know what I think about Borger's hypotheses, here you go:

The paper opens with an extended quote from Futuyma's evolutionary biology textbook on how mutations are random. Borger then covers various evidences of non-random mutations (what he calls "hot spots," and then he gets to the meat of the article that non-random mutation "provides an explanation for two phenomena associated with molecular phylogeny: homoplasy and nested hierarchy."

On homoplasy, he wrote,
Homoplasy is, in fact, just the illusion of common descent caused by inter-species hot spots recognized and acknowledged by evolutionists. How do evolutionists discriminate between random and non-random mutations when they are modelling the tree of life? In other words, how do they differentiate between homoplasy and real common descent mutations? The answer is they don't. It appears their trees are nothing but common design plus a handful of non-random mutations.

On nested hierarchy, he wrote,
The nested hierarchy is a reflection of DNA sequences that are more dissimilar in organisms that are more distantly related. This is what one might expect from common descent with modification; the longer the time since two organisms supposedly split from a common ancestor, the more differences should be observed. ...the nested hierarchy observed among species that cannot reproduce may just be a result of the functional restrictions (between separate designs) and non-random mutations. ... All functional domains and sites will contribute to the sequence identity of homologous genes in separate species. Phylogenetic analyses are therefore in effect a genetic mirage - the result of artificial constraints imposed by analysing functional domains together with non-random mutations - largely determined by the physico-chemical properties of the DNA sequence and its environment.

To someone with a background in molecular evolution (like myself), those quotes are pretty mind-boggling. I'm not really sure where to start, but let's try to look at homoplasy. Here's some purely hypothetical nucleotide sequences for three species:

cat   A G C T A
dog  A G C T G
cow G A C T A

In this example, we've got two invariant nucleotides and three that vary. Of the three that vary, two are identical in cat and dog (the first two) and one is identical in cow and cat (the last one). The simplest clustering would be to put the cat and dog together as the most similar with the cow as less similar. That's just basic clustering. An evolutionary interpretation would be that the dog and cat are closely related and the cow is a more distant relative. When we look at the last position where the cat and cow have an identical nucleotide and the dog differs, that would be considered a homoplasy in this dataset. As Venema pointed out, it could be caused by a novel mutation in the dog (meaning that the cow and cat have the ancestral sequence) or by independent mutations in the cat and cow that gave them the same nucleotide. Either way, in this dataset, it's a homoplasy - a nucleotide position that suggests a different clustering than the majority. The majority (the two nucleotide positions that exclude cow from the cat/dog group) determines the clustering, or "nested hierarchy," of these species.

With that in mind, let's assume for the moment that Borger is correct that hotspots contribute to the pattern of "common descent" (nested hierarchy). I think we can see that this sentence doesn't make sense:
Homoplasy is, in fact, just the illusion of common descent caused by inter-species hot spots recognized and acknowledged by evolutionists.
That can't be true because by definition homoplasy is not the data that gives you the pattern of common descent. If he was right that homoplasy is an illusion, there would be unambiguous nested hierarchies that would be interpreted as unequivocal evidence of common descent.

What about the nested hierarchy (the actual evidence of common ancestry)? The main thing I note from the above quote is that Borger's explanation is purely hypothetical:
...the nested hierarchy observed among species that cannot reproduce may just be a result of the functional restrictions (between separate designs) and non-random mutations.
(Note: I think by "species that cannot reproduce," he must mean species from different baramins.) The real key in this statement is "may." Unlike with homoplasies, he doesn't even try to provide evidence that non-homoplastic similarities (i.e., homologies) are caused by non-random mutations. He says in effect, "Maybe these nested hierarchy similarities are hotspots too!" That speculation doesn't remotely warrant his confidence that "Phylogenetic analyses are therefore in effect a genetic mirage." He's offered us no evidence to support that claim.

From a creationist perspective, I think this paper is especially frustrating since he seems to be assuming that differences in sequences are actually the result of mutation. Let's say for the sake of argument that he's right: that differences between chimps and humans are a consequence of non-random mutation. So that means that we started out with identical genomes? With the chimp, I suppose that's possible, but how far do we go with that before we have to admit that God created real genetic differences between species? Did humans start out with the same genome as gorillas? Orangutans? Monkeys? Mice? Chickens? Bacteria? Designed differences need to come into play somewhere, and Borger seems to acknowledge that when he wrote, "It appears their trees are nothing but common design plus a handful of non-random mutations."

The real question is how to determine whether an observed difference between two species is the result of post-Creation mutation or a created difference (i.e. common design). One simple way would be to look at differences that are polymorphic within the species. That would be an easy test of Borger's claim. If you can find evidence of intraspecific polymorphisms at the positions that Borger claims are hotspots, that would be good evidence that he's probably right. If not, then his claims are ad hoc.

As a creationist, I'm also convinced that there are non-random mutations but not at all for the reasons that Borger gives. If there really are baramins with many species all descended from a common ancestral pair (which I think there are), then the rapid production of those species virtually demands non-random changes, possibly even environmentally-induced changes. At this point, though, my conception of non-random mutation is also ad hoc, until someone can demonstrate a mechanism or a reproducible, adaptive change that is demonstrably not the result of selection.

As for the evidence of nested hierarchy and common descent, once we peel away this non-random mutation stuff (which Borger primarily equates with homoplasy), we're left with common design as explanation. As I've argued before (ad nauseum), without a model of design, "common design" is purely ad hoc. We're still left with the question, "If God really created, why did He make everything look like it evolved?"

In the end, then, I think Borger's "illusion of common descent" is just that: an illusion. Don't be fooled.

Feedback? Email me at toddcharleswood [at] gmail [dot] com.