This evening, I am watching an episode of that marvelous and profane Western, Deadwood, as I type this; it is a most excellently compensatory distraction, allowing me to sublimate my urge to express myself in uncompromisingly vulgar terms on Pharyngula. This is an essential coping mechanism.

I have been reading Jonathan Wells again.

If you’re familiar with Wells and with Deadwood, you know what I mean. You’ll just have to imagine that I am Al Swearingen, the brutal bar-owner who uses obscenities as if they were lyric poetry, while Wells is E.B. Farnum, the unctuous rodent who earns the contempt of every man who meets him. That imagination will have to hold you, because I’m going to restrain myself a bit; I’m afraid Wells would earn every earthy sobriquet I could imagine, but I’ll confine myself to the facts. They’re enough. The man completely misrepresents the results of a paper and a whole discipline, and does it baldly on the web, as if he doesn’t care that his dishonesty and ignorance leave a greasy, reeking trail behind him.

Let’s start with Wells’ own words.

Darwinists have been telling us for years that some of the best evidence for the common ancestry of insects and vertebrates is their Hox genes, which affect the character of body segments during embryo development. For example, a mutation in one Hox gene can cause a fruit fly to sprout a leg from its head in place of an antenna. Remarkably, vertebrates possess Hox genes that are very similar to a fruit fly’s — so similar that a mouse Hox gene can enable normal development in a fly embryo that lacks its corresponding Hox gene. More remarkably, the order in which Hox genes are lined up on the chromosome is the same as the order in which they’re expressed along the embryo’s body axis — a feature known as colinearity. And most remarkably, colinearity is the same in the four Hox gene complexes of vertebrates as it is in the Hox gene complex of the fruit fly Drosophila melanogaster.

This striking similarity in Hox gene colinearity is often cited as evidence for common ancestry. For example, according to a widely used college textbook, “The ordering of the genes within each vertebrate Hox complex is essentially the same as in the insect Hox complex, suggesting that all four vertebrate complexes originated by duplications of a single primordial complex and have preserved its basic organization.” (Bruce Alberts et al., Molecular Biology of the Cell, Fourth Edition, Garland Science 2002, p. 1194)

Last year, however, seven different arrangements of Hox genes were reported in various species of Drosophila — all fruit flies, in only one of the six model systems mentioned by Jenner and Wills. Apparently, the arrangement in Drosophila melanogaster that so strikingly resembles the arrangement in vertebrates has not been inherited from a common ancestor but is a relatively recent acquisition. (B. Negre and A. Ruiz, “HOM-C evolution in Drosophila: Is there need for Hox gene clustering?” Trends in Genetics 2006, doi:10.1016/j.tig.2006.12.001)

So this showcase piece of evidence for the common ancestry of animals – one of “evo-devo’s central themes” — turns out to be false, disproved by analysis of only one of the “big six” model systems.

Why, I’ve read the Negre and Ruiz paper, and know it well. I’ve also written previously on conservation of Hox structure, and am also familiar with that literature. Would you be at all surprised to learn that the major lessons Wells claims are in the paper, that fly Hox organization “has not been inherited from a common ancestor” and that “this showcase piece of evidence for the common ancestry of animals…turns out to be false” are conclusions not only unsupported by the evidence, but are contradicted by the work cited? No? You’re familiar with Mr Wells, then.

If you’re unfamiliar with what the Hox genes are, however, I’ve got a short primer on the subject. Even shorter than that article, I’ll mention that they are a set of genes with a specific DNA binding motif called a homeobox, and they are found in colinear clusters in the genome. That is, they tend to be arranged from 3′ to 5′ on the DNA strand in the same order that they are expressed from front to back in the animal; on one end in the fly, for instance, is a Hox gene called labial that is expressed at the front of the head, and at the other end of the cluster is one called abdominal-B that is expressed at the back of the abdomen. You don’t even need to know what the individual genes do, the interesting thing about them is their special organization on the chromosome.

What the Negre and Ruiz paper does is examine 13 species in the genus Drosophila (and one other dipteran, the mosquito Anopheles gambia), and reconstructs the evolutionary history of the Hox cluster, HOM-C. It is true that there are significant rearrangements in the Hox genes; this is old news, however. We knew even in Drosophila melanogaster that there were significant disruptions from the clean and tidy canonical order. One of the most obvious is that the fly cluster is actually broken in two, one complex called ANT-C and another called BX-C. Wouldn’t it be odd for developmental biologists to insist that strict and invariable colinearity was the sine qua non of homology when we knew from the first organism in which they were identified that their order was not absolute?

Here is the summary illustration of the data from that paper. Each of the genes in the cluster is marked with a thick arrow; dark blue are the Hox genes proper, light blue are Hox-derived genes (they are similar in sequence to the other Hox genes, but have derived and acquired new functions), and red are the non-Hox genes in this region. The similarities between all of the flies are relatively obvious, I think.

(click for larger image)

HOM-C structural evolution in the Drosophila genus. The Hox gene complex has suffered a high number of structural changes during the evolution of the
Drosophila genus. Three major rearrangements (shown as squares: A, Antp-Ubx; B, Ubx-abdA; and C lab-pb), seven microinversions (circles) and six gene transpositions
(inverted triangles) have been identified and mapped to the phylogenetic tree by comparative analysis. The structure of the HOM-C in each species has been analysed from
its complete genome sequence, except Drosophila buzzatii. Coloured arrows represent genes and their orientation; Hox genes are in dark blue, Hox-derived genes in
light blue and non-Hox genes in red. The Ccp gene cluster and the tRNAlys (denoted lys) cluster are not depicted in detail. (Subdivisions indicate the presence of several
genes, although not the exact number, which varies between species.) Double diagonal lines in the cluster diagrams represent discontinuities in the sequence, and yellow
shadows indicate equivalent breaks. The different segments are drawn in the order of the ancestral HOM-C and do not represent the actual order, orientation or distance
between chromosome segments. Crosses indicate different gene orientation in adjacent diagrams and have no phylogenetic meaning.

Ah, but look at the differences. The yellow bars indicate breaks in colinearity; the cluster has been broken up at those places and the pieces located in different regions of the genome. The X’s indicate inversions, a place where a stretch of the DNA has been flipped around relative to its neighbors.

The species D. simulans, D. sechellia, D. melanogaster, D. yakuba, and D. erecta are all closely related, and have the same Hox organization, with the genes in the same order and a break between the Antp and Ubx genes. Their more distant cousin, D. ananassae, has the same break, but also an inversion of the Dfd gene. Looking at even more distant relatives, flies that split from the melanogaster line over 60 million years ago, we see that D. grimshawi, D. virilis, D. mojavensis, and D. buzzatii do not have the Antp/Ubx split, but a different one, between Ubx and AbdA.

Actually, if you look closely at that diagram and puzzle out the rearrangements, something interesting pops out at you: the pattern of seemingly arbitrary changes in the arrangement of the genes fits neatly into a nested hierarchy, as if, for instance, the last common ancestor of D. melanogaster and D. willistoni had the Antp/Ubx split, and they just inherited the common pattern.

What? The evidence in the paper shows a pattern of inheritance of structure and variations from structure in the Hox genes? But didn’t Wells claim it showed that the arrangement wasn’t evidence of inheritance? How…odd.

You might have been tipped off that this was a story about evidence for evolution in a clade by the first four words of the title of the paper, “HOM-C evolution in Drosophila“. I suspect that the good Mr Wells may not have read that far.

There’s more. The paper is trying to explain the mechanism behind this slow pattern of changes in the Drosophila lineage, and it makes a good argument. The ancestral Hox pattern has a regulatory function; they are in that order because of a regional regulatory mechanism that switches the genes on in a timed sequence as the animal develops from anterior to posterior, that corresponds to their order on the chromosome. Flies, however, have a greatly accelerated rate of development. They have compressed the time of embryogenesis so much that a timing mechanism no longer works well, and instead they have evolved more complex individual regulatory elements for each gene. The regional control has eroded away, replaced by independent local control; as that has happened, the constraint that keeps them locked in a cluster has faded, and the genes have been free to drift apart.

Note that this is an evolutionary and developmental explanation for the phenomenon. It’s a successful account from an evo-devo research program, not a refutation of an evolutionary history. Mr Wells seems to have been wrong again.

I summarized a paper in 2004 that told this same story. It described the pattern of evolutionary change in Hox clusters in the chordate lineage, and we see exactly the same phenomenon, illustrated here.

Discrete changes of Hox gene complements in chordates. The chordate ancestor gained a rich set of posterior genes, which were inherited in the three subphyla but partly lost in ascidians. Central genes were gradually lost in tunicates, with larvaceans keeping anterior and posterior genes only. Whereas the Hox cluster was multiplied in vertebrates (with subsequent losses of a few paralogues in some clusters), the cluster degenerated in tunicates, and ultimately disappeared in larvaceans. The loss of central genes and of the Hox cluster coincides with a partition of Hox expression domains, which largely overlap in cephalochordates and vertebrates (ascidian data are still lacking). The motor of both events might be the decrease in size and transition to determinative development.

Animals that develop very rapidly and/or have very strictly defined (determinative) cell lineages in their embryos tend to evolve more discrete gene-by-gene regulatory control, lose the coarser control of whole blocks of genes, and the Hox clusters tend to disintegrate. One animal described in that work, the tunicate Oikopleura dioica, has completely lost the Hox clustered structure, scattering the individual Hox genes far and wide in the genome. Others have broken it up into several clusters.

Wells has a naive and incorrect view of Hox structure (it’s very peculiar; he was purportedly trained as a developmental biologist, and this ought to be common knowledge among the people I consider my peers and betters in this business. How he missed the basics, explained in multiple papers and books, I don’t know.) He seems to think the evidence for common descent of animals is a rigid core of unchanging DNA sequence, held in common in all animals. This is not true. What we have is roughly outlined common core, all right, but the evidence for evolution and common descent is a pattern of variation in that core Hox structure that fits a branching, lineal relationship.

In fact, I’d suggest that if we did see a fixed, invariant gene structure, free of variation between flies and frogs, it would be better evidence for design than evolution. Evolution is characterized by chance variation and interesting noise, except where it is constrained by selection, and the inheritance of contingencies in a lineage — and that’s exactly what we see in the paper Wells misleadingly claimed showed that common descent was false.

I really think that Jonathan Wells is … oh, wait. Al is delivering some choice words about Cy Tolliver. Just picture that, and you’ll have captured the essence of my opinion of Mr Wells, and I won’t have cluttered the old blog with such a ripe collection of pungent expletives.

Negre B, Ruiz A (2007) HOM-C evolution in Drosophila: is there a need for Hox gene clustering? Trends in Genetics 23(2):55-59.


  1. #1 Torbjörn Larsson
    April 4, 2007

    Then I passed out.

    Obviously drinking too deep from the Wells of lies isn’t good for us.

    But thanks for the splendid rebuttal on a complex subject!

  2. #2 Jonas Amore
    January 2, 2010

    To commercialize in terms of quality rather than price, and in order to differentiate accordingly, you require to adopt the standardised format of the 4 Ps marketing plan. That is, Price, Product, Place and Promotion manifestly you recognise the primary properties of the product, and the price, but for place you should think approximately the type of mass who are willing to pay over 4x price of competing product whereas the cheap option may be sold where accent is on cost, your ware will be suited to places/distributors where the clients will be willing to pay for high-performance. Thank you for this article! I’ve just detected a certainly open blog about how to get rich Test it!