Reading the entrails of chickens

Or, Reading the entrails of chickens: molecular timescales of evolution and the illusion of precision. Pointed out to me by a palaentologist friend. There’s a pdf here. Nothing at all to do with climate, but an interesting tale nonetheless. Or so I assume: it seems sensible, and was recommended by someone sensible, but may have been superceeded since 2004 for all I know. But this is the first time I’ve heard this wonderful story, and as someone who occaisionally reads about molecular clocks in the papers and assumes its all kosher, this article was a surprise.

Whats up (its fairly clear from the paper, I think, so I won’t bother summarise it more than briefly) is that estimates of species divergence from genetic clocks have been made with spurious accuracy and extrapolated well beyond reasonable bounds to produce erroneous results.

So we start with For almost a decade now, a team of molecular evolutionists has produced a plethora of seemingly precise molecular clock estimates for divergence events ranging from the speciation of cats and dogs to lineage separations that might have occurred ,4 billion years ago. Because the appearance of accuracy has an irresistible allure, non-specialists frequently treat these estimates as factual. In this article, we show that all of these divergence-time estimates were generated through improper methodology on the basis of a single calibration point that has been unjustly denuded of error. The illusion of precision was achieved mainly through the conversion of statistical estimates (which by definition possess standard errors, ranges and confidence intervals) into errorless numbers. By employing such techniques successively, the time estimates of even the most ancient divergence events were made to look deceptively precise. For example, on the basis of just 15 genes, the arthropod-nematode divergence event was ‘calculated’ to have occurred 1167 +/- 83 million years ago (i.e. within a 95% confidence interval of ,350 million years). Were calibration and derivation uncertainties taken into proper consideration, the 95% confidence interval would have turned out to be at least 40 times larger (~ 14.2 billion years).

And the language is wonderful: The reason for the transformation of the secondary calibration date into a primary one, which is equivalent to blood becoming Cabernet Sauvignon… and In what will surely not be the last chapter in this story, a recent review in Trends in Genetics [13] contains four blood-curdling innovations involving statistical methodology, taxonomy, physics of time reversal and logic.

Sigh. I wish climate papers were written like that.


  1. #1 RPM

    That’s one of my favorite papers because of the quality of writing. But it is also terrible how cruel it is. The same points could have been made without the vitriol and snark.

    In fact, the story behind its publication makes it even worse. Take a look at the references cited; they’re almost all papers by Kumar and Hedges. Usually when someone writes a scathing review of a particular scientist’s (or pair of scientists’) work, the scientists that are being attacked are given a copy of the review and a chance to reply in press. Kumar and Hedges were angry because of the tone of the Graur and Martin paper, but also because they were not given a chance to present their side of the story as a reply in the same issue of TIG. They eventually published their reply, but they weren’t very happy.

    Long story short, this review pissed off a few people. That said, it’s the best written molecular evolution article that I have seen. It’s too bad the writing is so mean spirited.

    [Hmmm yes, it lays into them pretty hard. From the outside, that made it more fun to read. I had assumed that they were assaulting the dominant paradigm, and that justified it -W]

  2. #2 John Sully

    I did note that it was published as an “Opinion” piece, still it was pretty fun to read. I take it that the error bars disprove evolution (ala M&M disproving AGW).

  3. #3 Hank Roberts

    With the error bars, they indeed do sound like they’ve disproved evolution. I dunno, snarky is so 20th Century a way of convincing people they are not only wrong but stupidly wrong.

    In matters of entrails, and error, I still prefer the approach to suggesting error used perhaps best by Oliver Cromwell after the storming of Drogheda, in 1649: “I BESEECH YOU IN THE BOWELS OF CHRIST THINK IT POSSIBLE YOU MAY BE MISTAKEN.”

  4. #4 Hank Roberts

    Hoist on own petard, I must admit my error in citation.
    Here’s the real source of the one quote, and the real quote with the correct source, along with the bits I left off.

    Cromwell, yes, but he was rather more unforgiving of error at Drogheda, and more so to the Scottish churchmen.


    * “I BESEECH YOU IN THE BOWELS OF CHRIST THINK IT POSSIBLE YOU MAY BE MISTAKEN.” — In a letter to the General Assembly of the Church of Scotland. 1650.

    I suppose having an army means he could be a boojum rather than snark.

  5. #5 SteveF

    The reply contains the following rather scathing comment:

    “Instead, they use histrionics, melodrama, hyperbole and an abundance of biblical references in their critique, while taking ad hoc approaches to expanding confidence intervals. This has only served to confuse and mislead those unfamiliar with the literature in this field.”

    Tee hee!

    Reminds me of a similarly pithy remark Doug Benn and Dave Evans made with regards to the megaflood hypothesis (not creationist related!) of glacial bedform development:

    “the idea of floods of such unimaginable dimensions is the outcome of taking flawed assumptions to their logical conclusion, a form of reductio ad absurdum in which the final absurdity is taken not as evidence of false premises but as fact”

  6. #6 Hank Roberts

    Today’s news:

    “… the poplar has four times as much DNA as Arabidopsis thaliana, the tiny cress which in 2000 became the first plant to have its genome cracked.
    “… By comparing the poplar information with that of Arabidopsis, the team has been able to show the two organisms split from a common plant ancestor about 100 to 120 million years ago….”

    Can someone comment whether the problems raised in the article first cited (“published as an “Opinion” piece” — has had any consequence for how work is done now? Or would those authors opine that today’s news is equally flawed?