S CIENCE’ S C OMPASS 65 text falls short of what might be expected 64 of an encyclopedia—the author’s lack of 63 expertise in several aspects of meteoritics 62 has resulted in errors. The straightforward 61 text takes the traditional approach of sys- 60 tematically describing stones, stony-irons, 59 and iron meteorites in terms of their clas- 58 sification and properties. The book is not 57 the sort that a nonspecialist would read in 56 order to learn about meteorites; it is much Image not 55 too technical for that. However, it is beau- available for 54 tifully and lavishly illustrated with many 53 pictures of meteorites in thin section as online use. 52 well as of hand specimens. These images 51 should help the work fulfill another of its 50 goals, providing a “guide to assist 49 searchers in the field to recognize the 48 many classes of meteorites.” 47 Meteoritics has come a long way in the 46 208 years since Chladni’s publication. Be- 45 van and de Laeter devote their final chapter 44 to looking to the future, in which they con- 43 sider missions to Mars, comets, and aster- BROWSINGS 42 oids, along with efforts to retrieve interstel- Earth from Above. Revised and Expanded Edition. Yann Arthus-Bertrand. Abrams, New 41 lar dust. Much remains for the next 200 York, 2002. 462 pp. $45.00. ISBN 0-8109-3495-7. Earth From Above. An exhibit at Mil- 40 years; we are still looking for meteorites lennium Park, Chicago, IL, through 30 September 2002. 39 that are indubitably from Mercury, Venus, a With a helicopter as his preferred tripod, Arthus-Bertrand specializes in composing im- 38 cometary nucleus, and a Kuiper belt object. ages on the fly. Ranging from an intimate glimpse at a worker resting on cotton in the 37 Part of the excitement of meteoritics is Côte d’Ivoire to a spectacular panorama of the Himalayan crest, the 190 photographs in 36 knowing that any of these extraterrestrial this oversize volume capture the beauty of natural landscapes and human settlements 35 rocks might already be here on Earth, just alike. Many depict strikingly abstract patterns, such as these formed by crystallized salts 34 waiting to be found, identified, and written on the surface of Kenya’s Lake Magadi (above). Arthus-Bertrand’s project was supported 33 about. In their different ways, both of these by UNESCO, and the book includes 14 new short essays on our beleaguered world and 32 books convey the message that meteoritics steps towards an “eco-economy.” The exhibit, with 120 large (1.2 m by 1.8 m) prints, 31 is a fast-moving field, one in which we still opened in Paris in May 2000 and has appeared in 40 cities and 15 countries. 30 have much to learn. 29 28 27 S CIENCE’ S C OMPASS PERSPECTIVES 26 25 PERSPECTIVES: GENOMICS the underlying evolutionary and genetic 24 mechanisms that shape them. 23 The tiger pufferfish is a good example of 22 Vertebrate Genomes Compared how the concept of “model organism” has 21 changed. Growing up to 70 cm in length, it 20 S. Blair Hedges and Sudhir Kumar is a relatively large marine fish known for its 19 taste, not a laboratory workhorse like the fly 18 t takes two of anything to make a com- account for this difference. Its unusually or mouse. It was introduced as a “genomic 17 parison. With the publication of the draft small genome size, combined with a faster model” organism (3) specifically because of 16 Igenome sequence of the tiger pufferfish method of sequencing (whole-genome its compact genome, permitting efficient 15 (Fugu rubripes) by Aparicio et al. on page shotgun), has yielded a much lower price comparison with the human genome. Other 14 1301 of this issue (1), we are now able to tag—a mere $12 million compared with the genomic models include human parasites, 13 compare the genomes of two vertebrates. hundreds of millions of dollars spent on se- such as Plasmodium and Trypanosoma, and 12 Measuring 365 million base pairs in length, quencing the human genome. The primary species of interest to agriculture, such as rice 11 the Fugu genome is only one-ninth the size incentive for sequencing this and other ver- and corn. Fugu’s relative, the spotted green 10 of the human genome (2) yet contains ap- tebrate genomes lies in better identification pufferfish (Tetraodon nigroviridis), is a 9 proximately the same number of genes. and characterization of human genes and much smaller (up to 17 cm in length) fresh- 8 Shorter introns and a smaller amount of their regulatory elements, especially those water species with a similarly small genome, 7 repetitive DNA in the pufferfish genome that are mutated in human diseases. For ex- and is more accessible to experimental re- 6 ample, nearly 1000 putative human genes search requiring laboratory breeding. The 5 S. Blair Hedges is at the NASA Astrobiology Institute have been discovered by comparing the Tetraodon genome, one of at least 18 verte- 4 and Department of Biology, 208 Mueller Laboratory, Fugu and human genomes (1). Beyond brate genomes being sequenced, is almost 3 Pennsylvania State University, University Park, PA 16802–5301, USA. S. Kumar is in the Department of biomedical applications, the extent of con- complete (see the figure). 2 Biology, Life Sciences 351, Arizona State University, servation and divergence among the puffer- A major motivation behind genome-se-

1 AGENCY,CREDIT:YANN ARTHUS-BERTRAND/ALTITUDE PARIS Tempe, AZ 85287–1501, USA. fish and human genomes will shed light on quencing projects is to generate a better

www.sciencemag.org SCIENCE VOL 297 23 AUGUST 2002 1283 S CIENCE’ S C OMPASS

65 Homo sapiens (Human) 64 Pan troglodytes (Common chimpanzee) 63 Macaca mulatta (Rhesus macaque) 62 Mus musculus (House mouse) 61 Rattus norvegicus (Norway rat) Canis familiaris (Domestic dog) 60 Felis catus (Domestic cat) 59 Equus caballus (Horse) 58 Sus scrofa (Domestic pig) 57 To marsupials Bos taurus (Domestic cattle) 56 Ovis aries (Domestic sheep) 55 Gallus gallus (Domestic fowl) Xenopus laevis (African clawed frog) 54 Xenopus tropicalis (Tropical clawed frog) 53 Danio rerio (Zebrafish) 52 Oryzias latipes (Japanese medaka) 51 Tetraodon nigroviridis (Spotted green pufferfish) 50 Takifugu rubripes (Tiger pufferfish or “Torafugu”) 49 Paleozoic Mesozoic Cen. 01234 48 Haploid genome 47 400 300 200 100 0 mass (pg) 46 Million years ago 45 Evolution of vertebrate genomes. The evolutionary tree shows relationships, times of divergence, and genome sizes (in picograms of DNA, pg) of 44 vertebrates whose genomes have been selected for sequencing. Classically, 1 pg of DNA has been considered equivalent to roughly 1 billion base pairs, 43 although the conversion factor for both human and Fugu is 0.91 ×109 base pairs. The position of marsupials is indicated to illustrate their potential im- 42 portance in filling an evolutionary gap in genome projects (12). The relationships and divergence times are largely from a molecular clock study of 658 41 proteins (8), although reflecting current uncertainty in the position of some orders of mammals (9) and supplemented with data on fishes and 40 genome sizes of vertebrates (www.genomesize.com/) (12–15). “Takifugu” is considered by ichthyologists to be the correct genus instead of the more 39 popular “Fugu”(www.fishbase.org/). 38 37 understanding of the genetics of human at evolutionarily conserved positions of the parts where times are unknown (8). Accu- 36 diseases (4). Sequencing of the complete protein sequence, and the alteration in chem- rate species divergence times are needed to 35 Fugu genome, a distant relative of humans, istry is more extreme than that permitted by link biological events with Earth’s geologi- 34 provides an opportunity to compare these natural selection (7). Thus, comparison of cal history and the appearance of charac- 33 two genomes at the level of exons, introns, nonmammalian genomes with the human teristic life forms. Such genomic clocks 32 and the protein sequence. The much short- genome will be crucial for understanding the might determine with better precision 31 er length of Fugu introns will enable se- genetic basis of many human diseases. whether the orders of mammals began 30 quence stretches important for regulating Evolutionary biology also has benefited splitting from one another in the Mesozoic 29 gene expression (5) to be identified, and greatly from genome-sequencing projects. era (80 to 100 million years ago) when 28 the role of noncoding DNA (which makes The wealth of new genome data is helping to continents were breaking apart, or later in 27 up most of the human genome) to be clari- better resolve the tree of life, particularly its the Cenozoic era (65 million years ago) af- 26 fied. The Fugu genome carries a handful of major branches. This has been especially true ter the dinosaurs disappeared, vacating 25 “giant” genes containing long introns, for prokaryotes, where more than 80 genomes ecological niches (8). Robust estimation of 24 which should provide insight into what has have been sequenced so far and the results the timing of gene-duplication events en- 23 driven the change in genome size during have greatly improved our view of the early ables reliable predictions of gene content 22 evolution. Understanding the genome history of life. For vertebrates, many thorny in unsequenced genomes and estimation of 21 structure of Fugu also will foster a clearer issues remain to be resolved, such as the phy- the rate of gene content increase over time 20 picture of the master control genes direct- logeny of families and other major groups in (11). Also, better estimates of when genes 19 ing vertebrate development (6). the tree of life. For example, it is not yet and species diverged will make it possible 18 The amino acid changes causing certain known whether humans are closer to mice or to investigate the relationship (if any) be- 17 human disorders are more likely to occur at to cattle because different results have been tween genome expansion (especially of de- 16 conserved (invariant) sites in the protein se- obtained with different gene analyses (8, 9). velopmentally important genes) and adap- 15 quence, and such sites are best identified by On the other hand, there is no guarantee that tation of species to their environments. 14 comparison with a distant species, such as complete genome sequences will immediately Historically, studies of biomedicine and 13 Fugu (7). In contrast, it is difficult to distin- solve all phylogenetic questions, as evidenced evolutionary biology have sought separate 12 guish functionally conserved sites in the pro- by the continuing debate over the relation- goals, yet comparative genomics (or more 11 tein sequence from less constrained sites ships among humans, flies, and nematodes precisely evolutionary genomics) is begin- 10 among close relatives, such as two species of (10). We will need to develop new statistical ning to blur traditional boundaries. A symbi- 9 mammal, because insufficient time has methods and bioinformatics tools to handle otic relationship has developed in which evo- 8 elapsed for changes to occur at the variable the greater volume of data and to unravel the lutionary biologists are answering fundamen- 7 sites. Why are amino acid changes in con- complexities of molecular evolution. tal questions using sequence data generated 6 served sites associated with human disease? Comparative genomics, with its large through public health projects. Simultaneous- 5 Sites that are less variable usually are more numbers of genes and proteins, provides ly, biomedical researchers are using trees, 4 crucial for survival of the organism, and better estimates of when different species time scales, and the methodology of evolu- 3 therefore changes at those sites are more diverged. These “molecular clocks” tell tionary biology to investigate human diseases 2 likely to cause disease. In fact, disease-asso- time by applying known rates of sequence and gene function. The future should see a

1 ciated amino acid changes are overabundant change from one part of the tree to other greater effort by the scientific community UNIVERSITY CREDIT: STATE RENEE GROTHE/ARIZONA

1284 23 AUGUST 2002 VOL 297 SCIENCE www.sciencemag.org S CIENCE’ S C OMPASS

65 and funding agencies to further recognize the Inherited Disease,C.R.Scriver, A. L. Beaudet, W. S. Sly, 10. J. E. Blair, K. Ikeo, T. Gojobori, S. B. Hedges, BMC Evol. 64 mutual benefits of comparative genomics. D. Valle, Eds. (McGraw-Hill, New York, 2000), vol. 1, Biol. 2,7 (2002). 63 pp. 259–298. 11. M. Lynch, J. S. Conery, Science 290, 1151 (2000). 5. B. S. Abrahams et al., Genomics 80, 45 (2002). 12. J. W. Thomas, J. W. Touchman, Trends Genet. 18, 104 62 References (2002). 6. M. S. Clark, L. Shaw, A. Kelly, P. Snell, G. Elgar, Im- 61 1. S. Aparicio et al., Science 297, 1301 (2002); published 13. C. A. Bisbee, M. A. Baker, A. C. Wilson, I. Hadji-Azimi, munogenetics 52, 174 (2001). 60 online 25 July 2002 (10.1126/science.1072104). M. Fischberg, Science 195, 785 (1977). 59 2. J. C.Venter et al., Science 291, 1304 (2001). 7. M. P. Miller, S. Kumar, Hum. Mol. Genet. 10, 2319 (2001). 14. B.Venkatesh, P. Gilligan, S. Brenner, FEBS Lett. 476,3 (2002). 58 3. S. Brenner et al., Nature 366, 265 (1993). 8. S. Kumar, S. B. Hedges, Nature 392, 917 (1998). 15. J. Wittbrodt, A. Shima, M. Schartl, Nature Rev. Genet. 9. W. J. Murphy et al., Science 294, 2348 (2001). 57 4. E. D. Green, in The Metabolic and Molecular Basis of 3, 53 (2001). 56 PERSPECTIVES: MATERIALS SCIENCE 55 perature, but with increasing temperature it 54 becomes faster than the frequency separa- 53 Dynamics in Ceramics tion of the peaks. The peak separation, typi- 52 cally 10 to 106 Hz, defines the time scale of 51 Jonathan F. Stebbins the site exchange (0.1 to 10–6 s). 50 Few studies of this kind have been per- 49 n a macroscopic scale, oxide ce- formed in inorganic materials, primarily 48 ramics are hard, rigid materials. It because ions of the most commonly ob- 47 is therefore tempting to think of 100.00 served nuclides (such as 29Si, 27Al, and O 17 46 them as rigid at the atomic scale as well. In 10.00 O) start hopping around at the required 45 particular, the oxide (O2–) ions that take up rates only at elevated temperatures. Alkali 44 much of the space in these materials seem- 4.00 metal cations are more mobile but often 43 ingly ought to stay put in the crystal lattice. 2.00 occupy only a single site in a given crystal 42 It is therefore a bit surprising that these 1.00 structure, also frustrating attempts at ex- 41 ions can be quite mobile in some ceramics. change spectroscopy, although there are 40 This mobility is exploited, for example, in 0.50 exceptions (5, 6). Fluoride anions can also 0.25 39 the zirconia-based (ZrO2) oxygen sensors be remarkably mobile, and high-resolution 19 38 that monitor the catalytic converters in 0.10 solid-state F NMR has recently revealed 37 most cars in the United States. The anionic F– site hopping (7). The fast ion conduc- 36 electrical conductivity supplied by oxide 0.00 tors described by Kim and Grey (1) are 35 ions is also crucial to many solid elec- Frequency unusual, with oxide site exchange rates 34 trolytes for fuel cells, which promise more rapid enough to be observed by NMR be- 33 efficient use of fossil fuel and hydrogen Merging peaks. Hypothetical NMR spectra for low about 250°C (the current upper limit 32 energy resources. a material that has two distinct sites with a 2:1 of fast magic angle spinning technology). 31 Detailed knowledge of the microscopic occupancy ratio. The spectra were calculated at In the ionic conductive ceramic α- 30 processes underlying ionic conductance is, increasing ratios of site exchange frequency to Bi4V2O11, Kim and Grey (1) report rapid 29 however, lacking. On page 1317 of this frequency separation (as labeled), as expected exchange of oxide ions among distinct 28 issue, Kim and Grey (1) bring a powerful with increasing temperature. sites within the vanadium oxide layers. 27 technique to bear on this problem. They The latter are rich in the oxygen vacancies 26 show that solid-state nuclear magnetic res- ions, it can readily be accomplished by that promote diffusive motion. No ex- 25 onance (NMR) not only can directly deter- high-temperature gas exchange (1). change is seen with the more rigid, fully 24 mine the rate at which oxide ions hop from Complementing the high-resolution struc- occupied bismuth oxide layers. In a relat- 23 site to site in ionic conductors, but also can tural data, NMR can also provide unique and ed, Ti-doped phase that is an even better 22 identify the sites involved in this process. quantitative insights into site-specific dynam- oxide ion conductor, the authors report 21 High-resolution NMR has long been an ics. Many such studies have been performed oxygen site hopping that is not frozen out 20 everyday tool for identifying organic on polymers (4), but site-specific dynamical until well below room temperature. 19 molecules in liquid solutions. Since the ear- data for inorganic materials remain limited. With some modifications, solid-state 18 ly 1980s, NMR has also been used widely Spin-lattice relaxation rates are most com- NMR can be used to observe exchange dy- 17 for structural studies of organic and inorgan- monly measured for nuclides such as 1H, 7Li, namics in a greater variety of materials. 16 ic solids. Solid-state NMR uses high-resolu- and 19F, whose ions often diffuse rapidly even Slower hopping rates can be observed by 15 tion methods such as magic angle spinning, in solids. Such studies can be informative but two-dimensional exchange experiments (5, 14 in which small samples are rapidly spun at are generally not site-specific. Furthermore, 7). Magic angle spinning NMR can be 13 an angle of 54.7° to the external magnetic data analysis can be highly model-dependent, done with special probes capable of reach- 12 field to average out some or all orientation- and relaxation can be strongly affected ing above 600°C, but with slower spinning 11 dependent spectral line broadening. For ox- by local “rattling” within sites, which is often rates and reduced resolution (6). Spectra 10 ide and silicate materials, 17O NMR pro- relatively unimportant to macroscopic can be acquired at temperatures above 9 vides a wealth of structural detail (2, 3) but transport properties. 2000°C on nonspinning (“static”) samples 8 requires enriched samples with ~1000 times The clearest sign of diffusive dynamics (8) and can show distinct sites and site ex- 7 their natural 17O abundance. Enrichment in NMR spectra is the merging of two spec- change for nuclides such as 27Al, 29Si, and 6 sometimes necessitates tricky syntheses, but tral peaks, assigned to ions or molecules in 17O (9). Static spectra of single crystals of- 5 in refractory materials with mobile oxide distinct chemical environments, with in- ten have high resolution and can again be 4 creasing temperature (see the figure). The collected at high temperatures (10). 3 merging process or “coalescence” is often The prospects are thus bright for eluci- The author is in the Department of Geological and En- 2 vironmental Sciences, Stanford University, Stanford, attributed to chemical exchange between the dating mechanisms of ionic mobility in 1 CA 94025–2115, USA. E-mail: [email protected] sites. The exchange rate is slow at low tem- oxide materials with high-resolution

www.sciencemag.org SCIENCE VOL 297 23 AUGUST 2002 1285