American Scientist

The Delphic Boat: What Genomes Tell Us. Antoine Danchin. Translated by Alison Quayle. x + 368 pp. Harvard University Press, 2002 (originally published as La barque de Delphes [Editions Odile Jacob, 1998]). $35.

Years ago in Paris, I chanced on a small gem of a science museum. Displayed in the suitably hushed space was much of the 19th-century history of physics. Although the scientific apparatus was familiar, the names of the people affixed to the various pieces of equipment were not. I had stumbled into an essentially parallel universe in which a surprising portion of modern science was considered French in origin. Apparently, in science, as in most things, there is a distinct Gallic sensibility.

The Delphic Boat, by Antoine Danchin, reflects, I suggest, some very French attitudes, including a yearning for global principles and a deep, sometimes insufficiently critical, appreciation for originality. The title refers to an old conundrum: If all the planks in a boat are replaced, does it remain the original vessel? The answer is that "boat-ness" lies in the relation between the parts, not necessarily their identity.

Danchin posits that because a number of genomes have been sequenced and most of the gene products of living cells have been identified, biology must now consider their interrelations. The point is well taken—much of modern molecular biology remains bound to its origins in the architectural formlessness of solution biochemistry, with only tentative studies of the organization of higher organisms. Alas, Danchin's own experience seems limited to bacteria (he heads the Unit of Genetics of Bacterial Genomes at the Institut Pasteur). He also is fond of metaphors, uses scientific terms loosely (a practice that, ironically, he inveighs against) and displays an ill-considered willingness to do battle with even the most solidly established fields of physics and information theory. He does not, I believe, offer a useful guide to future thought.

The book begins with a lively, gossipy insider's description of the genome battles. Like most scientific gossip, it is of considerable, if slightly guilty, interest. Inevitably, the triumphalism of genome sequencers soon appears—the view that this great scientific achievement is an open sesame to the nature of life and other deep questions. Danchin assumes that since we are now in a position to know in detail the nature of all "genes," the correct approach will yield complete understanding of biology. This of course is predicated on the widely held—but, I believe, unsupportable—assumption that genes do little else but specify proteins. Usually ignored is the major portion of metazoan genomes that is vigorously transcribed but, since it produces no open reading frames (possible protein-coding regions), is variously dismissed as "junk"—or, in the author's ingenious (if highly unlikely) suggestion, a repository for discarded sequences that make information transfer reversible, at least in principle. This rather far-out suggestion arises from Danchin's insistence that previous estimates of the minimal energy for computation are all wrong.

The more philosophical, speculative and, indeed, poetic products of Danchin's thought are in the chapters that follow. For example, after admonishing the reader to beware the misuse of words, he strides fearlessly onto tremulous ground. Thermodynamics, entropy and, distantly, information theory seem irresistible topics. The concept of energy, which he curiously ascribes to Mesmer, becomes conflated with social discontents. According to Danchin, work defines the useful part of energy, and the remainder, entropy, is considered degenerate and even dangerous. In physics class, work is defined simply as force multiplied by distance, while entropy accounts for why some steam engines consumed more coal than others for the same output. Indeed, the danger of verbal insight is well illustrated by Danchin's disagreement with the conclusion of statistical physics that the homogeneous state is the most disordered. Even though a trampled sandbox seems to him disorganized compared with a smoothly raked one, actually, in principle, work can be derived from the disheveled sandbox, which is therefore in an entropically lower, more ordered state. I suppose that his identification of entropy as "more a measure of exploration than of anything else" attempts to convey what physicists refer to as a system's diffusion through all possible configurations.

Danchin vigorously and, I believe, correctly decries the "tricks of language and abuse of metaphor" used to invoke self-organization as an explanation of the complexities of biology. Unfortunately, he then goes on to posit "symbolic abstraction" as the basis of life. I must admit utter bafflement at this point. Danchin views the genetic code as symbolic, metaphoric and so on. Surely, it is about as symbolic as a camshaft. The book repeatedly points out that the mysterious principles of life are not predicted by physics, as in Schrödinger's equation—but, then, neither is the weather. About the latter nothing is mysterious save its complexity.

Danchin ranges over a vast spectrum of ideas, from the pre-Socratic philosophers through Immanuel Kant and Samuel Pepys to the latest information theorists. There's no denying his erudition and wide interests; he is probably a superb dinner conversationalist. The book does note a number of nagging biological questions that are currently swept under the rug, but regrettably it does not resolve any of these.—Sheldon Penman, Biology (Emeritus), Massachusetts Institute of Technology