Friday, October 19, 2007

The mystery in the comments

Derek Atkins (apparently my only reader these days) has forced me out of blog inaction by a series of comments to an old piece on Sanford's curiously titled book. You may recall that I started reading it out of politeness after Salvador Cordova gave it to me. I quickly lost interest once I realized just how bad it was and decided to spare you a complete series of negative reviews. After all, I have serious work to do writing papers and grant applications, going to seminars, pushing worms (they've missed me!), analyzing data, mentoring and teaching. Somehow I doubt that my tenure package will look stronger if I waste valuable time refuting bad (really, really bad; failing student bad) population genetic arguments -- a subject which, by the way, I'm teaching right now to a dozen (suffering) undergraduate and graduate students.

Now Atkins has bombarded me with so many comments that I may never get around to answering all of them, but I see a few that are worth responding to, because they are representative of the kinds of arguments that Sanford and other creationists make routinely. After a discussion on how many different kinds of creationists can dance on the head of a pin, Derek gets to what is really bothering him: "What I take exception to is your list of Sanford’s assertions that twists what he really does say. This speaks to your integrity." So I'm in trouble now... Here goes:

[Continue reading below the fold.]

Let’s start with the first assertion to which you’ve taken exception. Analogies are what they are—an attempt to aid the reader (since there may be a wide range of exposure to the materials) in understanding the salient points. It is a device that is effective and is used in all disciplines. [...] The use of differing language/information analogies to the genome does not start nor end with YE or ID proponents (they are not one-in-the-same) but is seen in the analogies of those who are strident evolutionists as well, such as Carl Sagan (1974) when he compares the genome with having more information than contained in the Library of Congress. [...] What I find interesting is not having read any evolutionist being singled out for using these analogies, but anyone who is YE or ID is denigrated for doing so. Unless you have taken non-YE/IDers to task regarding this point, it seems irrelevant to single out Sanford. I credit Sanford for being quite clear about the limitations of the analogy, as quoted above.
Yes, Dawkins used the cake recipe in 1982 in The Extended Phenotype, Dennett used Borges' Library of Babel concept in an interesting analogy for genotypic space in Darwin's Dangerous Idea, and Steve Jones wrote a whole book on The Language of the Genes. Good science writers make their careers on good analogies. Of course to suggest that they never get into trouble over their analogizing is ridiculous: entire volumes of criticism by scientists and philosophers have been devoted to the "gene/meme" and "spandrels" analogies alone. The problem is that, despite his mild disclaimers, the analogy Sandford introduces is a terrible one to discuss the evolution of genomic complexity. So I pointed it out. The reason I did this without hesitation and with what you might perceive as some impatience is that creationists of all stripes use analogies as substitutes for the rigorous scientific work of formulating hypotheses, testing them in the laboratory, or with computer simulations, or with observations from nature, and developing theories based on these activities.

Moving on:
In the second argument you offer as one example the nematode Caenorhabditis elegans, citing de Bono M, Barmann CI,.(1998). In the abstract we read, “A loss-of-function mutation in the npr-1 gene, which encodes a predicted G protein-coupled receptor similar to neuropeptide Y receptors, causes a solitary strain to take on social behavior.” [...] Loss-of-function and various other degradations to the genome is more the rule than the exception to genetic mutations. This is the point Sanford makes clear, though you sidestep this and ask, “Does it matter?” It is an important area of discussion that is not well served with a flippant response. Any gene that mutates causing loss-of-function has certainly lost, not gained usefulness, and by definition the genome has lost information."
This is wrong at so many levels that it is even hard to know where to begin. First of all, it is by no means certain that the genome has lost any information in this case (or any other involving mutations). The main reason is that Sanford and other people who use this argument don't actually define the term in a meaninful way (precisely one of my main criticisms of Sanford). According to any technical meaning of information that I am aware of, a single aminoacid substitution does very little, if anything to it.

The second error you (both) make is that you misunderstand what is meant by the expression "loss-of-function". This has nothing whatsoever to do with information. It also does not mean degradation. de Bono & Bargmann are referring only to the activity of the npr-1 gene product, which is a transmembrane receptor that binds to small peptides outside the cell and creates a signal inside the cell. Loss of function in these cases means that the peptides bind with lower affinity to the receptor or elicit a weaker signal inside the cell. The "social allele" is a loss of function allele when you treat the "solitary allele" of the N2 strain as the reference allele (the reference strain in C. elegans biology). If the Hawaiian strain CB4856 happened to be the standard strain, instead of the English N2 strain, the mutation would be referred to as a gain-of-function mutation. Note that gain-of-function mutations in all kinds of genes are routinely isolated by geneticists (something Sanford neglects to mention). What de Bono & Bargmann are saying is that the NPR-1 receptor from a social worm has a lower activity than the receptor from a solitary worm. But if you look at the level of the behavior, it is not obvious which one has the greater "function": they are both functional, they just have different functions. The irony is that the best information available suggests that the social allele (with lower activity) is actually ancestral to the solitary one in the Caenorhabditis genus, so the mutation most likely occurred in the direction of increased function in the evolutionary history of C. elegans.

So let's recapitulate. On the first point,
you did not actually address my substantive, specific criticisms of his analogy. To say that others have also used analogies is beside the point. The question is: where exactly did I twist what Sanford was saying? On the second, until you define information, Sanford's point remains nonsensical. Your misunderstanding of the meaning of loss-of-function does not inspire confidence in your command of the issues at hand. Do yourself a favor and read a population genetics textbook before embarrassing yourself any further: Gillespie's Population Genetics or Hartl & Clark'a Principles of Population Genetics are both pretty good.

Read on