Modern Phylogeneticists Branch Out

Courtesy of Andrew Syred, Science Photo Library COGS IN THE DATABASE: Duke University's Mitchell Levesque knocked out flagellar function in Bacillus subtilis, like those shown above, to show that novel COGs [clusters of orthologous groups of proteins] can be identified using a single trait-to-COG approach. For nearly a century, biologists relied on fossil records and morphological comparisons to reveal the evolutionary histories of organisms. But most phylogenetic trees need more than phenotypic evidence to weather the intense debate in which reconstructionists are known to engage. With the advent of molecular phylogenetics in the 1960s came the hope of better trees, but it took nearly 40 years for them to materialize. With sequencing and analysis tools, modern phylogeneticists are finally building trees that can withstand scientific scrutiny. This issue's first two Hot Papers demonstrate DNA's ability to reveal novel branching patterns in mammalian evolution.1,2 A third discusses using DNA sequence changes across evolutionary history to extrapolate function and even study genomic evolution.3 Researchers working to build definitive trees are finding their hopes fulfilled, and then some (see 5-Prime p. 14). With the largest DNA sequencing database ever used to determine placental- mammal evolution, scientists from Stephen O'Brien's lab at the National Cancer Institute (NCI) reconstructed the branching patterns of the 18 extant placental orders. Previous morphological evidence had not even hinted that these 18 orders were divided among the four major groups, or superorders, which the DNA sequence revealed.1 Another team from Mark Springer's lab at the University of California, Riverside, independently came to the same conclusions.2 The two later collaborated, using the database to show when these orders started diverging into their respective groups.4 Data derived from the Science Watch/Hot Papers database and the Web of Science (ISI, Philadelphia) show that Hot Papers are cited 50 to 100 times more often than the average paper of the same type and age. W.J. Murphy et al., "Molecular phylogenetics and the origins of placental mammals," Nature, 400:614-8, Feb. 1, 2001. (Cited in 127 papers) O. Madsen et al., "Parallel adaptive radiations in two major clades of placental mammals," Nature, 400:610-4, Feb. 1, 2001. (Cited in 87 papers) R.L. Tatusov et al., "The COG database: New developments in phylo-genetic classification of proteins from complete genomes," Nucleic Acids Res, 29:22-8, Jan. 1, 2001. (Cited in 152 papers) Also taking advantage of the phylogenetic power of DNA, a group of computational biologists, led by Eugene Koonin at the National Center for Biotechnology Information, is using DNA sequence data to construct clusters of orthologous groups of proteins (COGs). These protein groups from different species could be related through evolution. Usually, researchers access the publicly available COGs database to extract functionalinformation about particular protein sequences. But some, including Koonin, use it as a jumping-off point for studying genome evolution. Koonin and his collaborators have continually updated the database since its inception in 1997 and describe one of its recent incarnations in the third Hot Paper.3 A NEW WORLD SUPERORDER Traditional phylogenetic approaches could not resolve the evolutionary history of placental mammals, which make up almost 90% of the approximately 4,600 mammalian species that exist today. Placental mammals are believed to have diverged from marsupials about 170 million years ago, but it is unclear what happened afterward. Some paleontologists say today's 18 orders started diverging before the Cretaceous- Tertiary (K-T) boundary an estimated 65 million years ago; others say it happened later. "The phylogeny of mammals has been a controversy for a very long time," says the NCI's Bill Murphy.1 Murphy, Eduardo Elzirik, and their collaborators described how they sequenced 15 nuclear and three mitochondrial genes, totaling nearly 10,000 base pairs, from 64 species sampled across all 18 orders of extant placentals. Five years ago, Murphy explains, molecular phylogenists who were studying these problems were basing their trees on single genes--usually about 1,000 base pairs--which generally don't provide much resolution. Springer and his colleagues sampled two independent molecular data sets; one totaled nearly 6,000 base pairs, and the other contained 3,000. Both groups independently postulated that placentals comprise four major evolutionary groups. Previous morphological evidence had predicted the existence of only one of these so-called superordinal clades. The others were a "major surprise," says Springer. The superorders, monophyletic groups of organisms more closely related to each other, are: Afrotheria, including elephants; Xenarthra, including armadillos and anteaters; Laurasiatheria, including carnivores and bats; Euarchontoglires, including humans and rodents. The data also resolved some of the smaller branching patterns within each major clade; for example, determining that megabats and microbats are monophyletic. But, it was not defined whether the superordinal clades had diverged before or after the K-T boundary. After publishing, the two teams joined forces and, using a combined data set, showed that the major groups started diverging about 103 million years ago when the superordinal Afrotherians split from the other placentals. This division coincides with the break-up between Africa and South America. The other three clades also split before the K-T boundary.4 After another combined effort using a more robust methodology, the scientists finally built a placental mammalian tree that stands a good chance of weathering the storm of phylogenetic debate.5 But the work isn't done. Says Murphy: "We have almost completely resolved the placental mammal tree and resolved the controversy about when primordial mammals diversified." Their results corroborate the "Long Fuse Model," one of the three competing hypotheses for placental mammal diversification, which states that interordinal divergence occurred before the K-T boundary, and intraordinal divergence occurred mostly afterward. Courtesy of J. Pettigrew TIGHTER TIES: The so-called flying primate hypothesis states that the primate-like megabats, or Old World fruit bats (bottom image) evolved independently from the microbats, or echolocating bats. Not so: The new data assert that they are monophyletic. SHARING THE WEALTH The new tree isn't just for mammalogists. Genomics researchers, such as Eric Green at the National Human Genome Research Institute (NHGRI) are using the data to make informed decisions about which species would be most useful for annotating the human genome. Currently, only the mouse and rat genomes can be used as reference points. As it turns out, rodents and humans are in the same superorder, Euarchontoglires. Using genomes from one of the other three superorders likely would allow for much more powerful analyses. Cows, dogs, and pigs are three non-Euarchontoglires placentals on NHGRI's priority list of model organisms to be considered for complete sequencing, but neither of the other two major clades is represented. Escherichia coli, arguably the best-studied species on Earth, only has had 40% of its gene products characterized, demonstrating that there is ample room for annotation improvement, even for prokaryotes. For this and other reasons, Koonin and his collaborators developed the COGs database, one of several new computational tools that researchers can use to extract functional information about unknown protein sequences. A COG is a group of three or more identical protein sequences from different species that are believed to have the same evolutionary history. Members of any given COG usually have the same function, no matter what the species. So, by identifying which COG an unknown protein sequence belongs to, explains Koonin, one can make an "educated guess" about its function. When the database first went public, it contained 720 COGs from six prokaryotes and one eukaryote, Saccharomyces cerevisiae. By the time the Hot Paper was published, the database held 2,791 COGs from 30 prokaryotic genomes and the yeast. Since then, it has grown to about 5,000 COGs distributed among 65 prokaryotes and S. cerevisiae, and another 6,000 COGs from seven additional eukaryotic genomes, including the human. As the database has grown, so too has its annotative potential. Graduate student Mitchell Levesque from Duke University has designed algorithms that identify which COGs are involved with any given function. Levesque explains that if a group of different species has all retained a certain trait and their genomes all have the same COG, then that COG probably has something to do with that trait. Levesque and his collaborators used knockout Bacillus subtilis to show that even novel COGs--in this case, COGs without any previously known flagellar function--can be identified using this trait-to-COG approach.6 Beyond annotation, Koonin maintains the database primarily to springboard his own work on genomic evolution. COGs that reveal unusual phylogenetic patterns are usually the first clue that evolution has taken an unexpected turn, Koonin says. If, for example, a COG that exists in all eukaryotic species occurs in only one bacterial species, then the bacteria may have acquired the gene by horizontal transfer from another species. Some scientists say that horizontal gene transfer may not be so unusual after all. In fact, it may occur so frequently that it threatens to uproot the very notion of a phylogenetic tree, at least among prokaryotes.7 But for the moment, the placental tree still stands strong. Leslie Pray (lpray@nasw.org) is a freelance writer in Northampton, Mass. References 1. W.J. Murphy et al., "Molecular phylogenetics and the origins of placental mammals," Nature, 400:614-8, 2001. (Cited in 127 papers) 2. O. Madsen et al., "Parallel adaptive radiations in two major clades of placental mammals," Nature, 400:610-4, 2001. (Cited in 87 papers) 3. R.L. Tatusov et al., "The COG database: New developments in phylogenetic classification of proteins from complete genomes," Nucleic Acids Res, 29:22-8, 2001. (Cited in 152 papers) 4. W.J. Murphy et al., "Resolution of the early placental mammal radiation using Bayesian phylogenetics," Science, 294:2348-51, 2001. 5. M.S. Springer et al., "Placental mammal diversification and the Cretaceous-Tertiary boundary," Proc Natl Acad Sci, 100:1056-61, Feb. 4, 2003. 6. M. Levesque et al., "A computational method for predicting the function of uncharacterized genes," Curr Biol, 13:129-33, Jan. 21, 2003. 7. W.F. Doolitle, "Phylogenetic classification and the universal tree," Science, 284:2124-9, 1999. function sendData() { document.frm.pathName.value = location.pathname; result = false if (document.frm.score[0].checked) result = true; if (document.frm.score[1].checked) result = true; if (document.frm.score[2].checked) result = true; if (document.frm.score[3].checked) result = true; if (document.frm.score[4].checked) result = true; if (!result) alert("Please select") return result; } .myradio{background-color : #94bee5} Please indicate on a 1 - 5 scale how strongly you would recommend this article to your colleagues? Not recommended 1 2 3 4 5 Highly recommended Please register your vote

Modern Phylogeneticists Branch Out

Interested in reading more?

Become a Member of

Courtesy of Andrew Syred, Science Photo Library
COGS IN THE DATABASE: Duke University's Mitchell Levesque knocked out flagellar function in Bacillus subtilis, like those shown above, to show that novel COGs [clusters of orthologous groups of proteins] can be identified using a single trait-to-COG approach.