Friday, September 19, 2008

Homology: what's evolution got to do with it?

British palaeontologist Colin Patterson became an unwitting friend of creationism during his career. That misbegotten legacy continues to this day, in misquotations that continue to pop up in creationist literature. Patterson has been widely cited by creationists as some sort of closet creationist who though evolution was a speculative farce. Unfortunately, what has become lost in the maelstrom of attack and counter-attack in the world of creation/evolution apologetics are lessons for both evolutionists and creationists.

If anybody was a skeptic, Patterson was. In the forward to his posthumous second edition of his textbook Evolution, two of his close colleagues wrote: "His favourite critical internalised question was 'how do we know that?': to which he often got the answer 'authority or tradition'--something he respected only after he had explored the evidence for himself". He was notorious for a need to figure things out for himself and endeavoured like no other to never let preconceptions get in the way. Indeed, based on anecdotes of people who knew him, I have learned that "how do you know that?" was not merely internalised, but frequently vocalised in a deep Oxford English from the back of the room.

As a result of Patterson's take-no-prisoners approach to belief and science, he became the champion of some unpopular ideas. Patterson questioned every authority and, in the end, challenged (and I believe overturned) some deeply held beliefs about evolution. More importantly, he overturned some ideas about how we know what we know about evolution. For instance, even as a palaeontologist, he argued strongly that fossils themselves play little (if any) role in the establishment of species relationships. That belief emerged from beliefs about fossils revealing ancestor-descendent relationships, and from prior commitments about transformation.

This is where the creationist and evolutionist misunderstandings commence. Patterson argued that evolutionary theory had no role to play in systematics. To creationists, this is touted as evidence that the theory of evolution has no practical applications and is, indeed, unnecessary in biology. To evolutionists, this is often either ignored, disagreed with, or misunderstood.

But what Patterson showed was that a lot of the pre-Darwinian basis for evolutionary theory had been co-opted or subsumed into evolutionary theory. Ideas that had a pre-evolutionary basis had become drenched in evolutionary pre-conceptions and language. Homology, for instance, had become (and still is for most): shared similarity due to common ancestry. However, if homology is explained by common ancestry, then what is the basis for the inference of common ancestry? Well, as it turns out, homology! Patterson recognized the problem and iterated a definition of homology that took into account the way in which homologies define nested groupings. That is, homology is the relation that defines the ranks in a nested hierarchy.

Homologies are homologies because they define nested groups. They are sets of characters that fall into a series of congruent groupings. Similarity alone isn't enough to justify statements of homology, otherwise, we have no way to distinguish convergence from homology. Many will cite examples such as bird, bad, and pterosaur wings as examples where "fundamental differences" allow us to distinguish homology from non-homology.

But, the reality is that we already know these structures are non-homologous because they appear in distantly related groups. We know they're not homologous because of the distribution of other characters which act as a test of homology. If there were ample character evidence that birds, pterosaurs, and bats were all a tightly related group, we might then explain the differences as specializations of a common ancestral wing. The test, ultimately, is whether these taxa share other important characters in common.

The consequence of Patterson's definition of homology--the relation that defines a monophyletic group--is that evolutionary preconceptions are not necessary. Many evolutionists are uncomfortable with this. There is a sort of pluralistic approach (what I call a 'holistic' approach) to homology assessment that many biologists subscribe to. People argue that as many lines of evidence as possible should be considered. I agree, but the question is, through what filter do we analyze this evidence? Patterson would have answered: "tradition, authority, convenience, or assumptions about evolution". There is actually no need to be uncomfortable with Patterson's approach, which I'm surprised has not become more widely embraced.

The problem, as others had pointed out before Patterson, was that we need a knowledge of phylogeny (or interrelationships) in order to know anything about evolutionary history. In order to make generalizations about how evolution works, we need to know the pattern of descent. However, if our assumptions about evolution feed into our inferences about the pattern of descent, these assumptions become untestable. As a result, Patterson argued that our beliefs about evolution played no role in systematics. It was the task of systematists to uncover the patterns that exist in nature which we choose to explain by evolution and common descent. Patterson's rejection of the role of evolutionary theory in systematics was an attempt to keep the enterprise from decaying into circular argumentation.

So, the lesson for evolutionists should be kept in mind as we are deep into a new age in comparative biology. Genes and proteins can now be sequenced, we can map gene expression to embryos, and study the fate of populations of cells in developing embryos. We must ask ourselves: what beliefs about evolution that we developed before these wonderful advances have we carried with us to the present? And for each of these beliefs have we asked: how do we know that?.
Read full post

Thursday, September 18, 2008

Open thread: are genes really a guide to homology?

I have been putting this question to some of my colleagues:

What is the value of gene expression data in determining homology of morphological features?

Are genes really important in determining if two structures in two different animals are homologous? If so, why? If not, then what does really matter?

Read full post