by J. William Bell, Access Magazine Editor -- "There's been about 94 million years of evolution since cattle and primates diverged from a common ancestor, so it doesn't take a doctor to recognize the difference," says Harris Lewin, director of the University of Illinois at Urbana-Champaign's Institute for Genomic Biology and a professor of immunogenetics. It does, however, take a lot of work and specialized tools to relate species' genomes to one another and to uncover the finer points of distinction and similarity. Lewin's team at the U of I recently published the most extensive comparison of mammalian genomes ever created--looking at the chromosome organization of humans, cattle, rats, mice, cats, dogs, horses, and pigs, all at once. They collaborated with researchers at Texas A&M University, the Genome Institute of Singapore, the National Cancer Institute, the University of California at San Diego, and other institutions. Results, presented in a wide-ranging Science article in July 2005, showed that the historical rate of chromosome evolution in mammals was different than previously thought. They also revealed provocative new features of chromosome breakpoints in evolution and cancer.
The comparisons relied on Evolution Highway, data visualization and analysis software built for the team by NCSA. It is based on NCSA's application-development environment for data mining called D2K.
Dare to compare
To begin their work, the Lewin team collected genetic maps and sequences for each species. Research teams around the world provided data, including the Lewin team itself, which constructed the cattle map in collaboration with James Womack's group at Texas A&M.
Maps and sequences both yield a run down of what genes and other DNA sequences appear on which chromosomes. The differences come in the level of detail. In a whole-genome sequence, all or nearly all base pairs, which number in the billions, are decoded. The positions of most genes are identified in the sequence. To construct a genetic map, particular genes and other DNA sequences are selected, isolated, and ordered on the chromosome using a variety of approaches. Mapping gives a less precise but more easily gathered result.
The resolutions of the gene maps of each species are improving, and new gene sequencing efforts are constantly under way. But the uneven distribution of genes over the genome and the uncoordinated way that far-flung groups select genes for mapping complicate comparisons of one mapped species to another, according to Denis Larkin, a visiting assistant professor in the U of I's animal sciences department.
Instead, the Lewin team used an algorithm called BLASTn to compare millions of DNA sequences from each species' genome to the complete human sequence. More than a thousand homologous synteny blocks were identified. These blocks are segments with only minor differences, marked by tiny rearrangements in the base pair order rather than prolonged inconsistencies. "We find those areas that are most similar statistically, not necessarily exact matches," Annelie Everts-van der Wind, a PhD candidate on the team, points out.
Frequently, homologous blocks shift about, even when located on the same chromosome. The same genes might be found on chromosome 17 in humans and chromosome 19 in cattle but the homologous blocks of genes may be shuffled, for example. Within blocks, which may be as large as 50 million base pairs of DNA, gene order may be preserved for hundreds of millions of years. The predictability of the comparative maps allowed the team to use COMPASS, a software package developed by the Lewin group in 2000, to predict the positions of homologous blocks in a species when their positions in the human sequence were known.
'Deep biological significance'
The genes we have in common with other species are important. They allow researchers to identify what's responsible for the basic regulatory functions that all biological systems share, among many other things. But places where deviation occurs--breakpoints in the compared maps that flag fission, inversion, or other changes in the code--also tell a story.
"The core hypothesis is that breakpoints have deep biological significance," Lewin says. The Science article focused on those breakpoints and that hypothesis.
First, the team's eight-species comparison allowed them to draw some general conclusions about the rate of mammalian evolution. Looking at cross-species comparisons using the NCSA-built Evolution Highway software, they determined the number of breakpoints that appeared in a given species but not another particular species. They then divided this number of unique breakpoints by the number of years since the two species diverged. Estimates on the number of years were based on other researchers' studies of the fossil record, as well as morphological and genetic studies.
"For example, if there are 67 breakpoints specific for cattle but not found in pig, and [outside research tells us that] pig and cattle diverged about 60 million years ago, then the cattle genome evolved at a rate of 67 divided by 60 or 1.1 chromosomal breaks per million years on average," Larkin explains. Meanwhile, if "there are 10 breakpoints common for cattle and pig, we can assume that those formed in a cattle-pig ancestor after divergence from an ancestor they shared with cat and dog and before the cattle-pig split. So if that ancestor existed for 25 million years and accumulated 10 breakpoints, it evolved at a rate of .4 breaks per million years."
Early works on genome comparison held that evolution moved at a more rapid pace between 100 million and 65 million years ago and has since slowed. The team's study broke with the conventional wisdom, showing that evolution has been moving faster for the last 65 million years than it did for the 35 or so million years before that. They're confident in their numbers because older studies relied on much more rough genome maps and comparisons among fewer species.
The team looked at more than just the aggregate number of breakpoints, though. They also considered "reuse" breakpoints, which show up in the same place across multiple, disparate species and are thought to mark fragile places in the genome where rearrangements are more likely to occur.
Early genomic studies in the '80s and '90s typically ascribed to a random breakage theory. Recent research, including a comparison between the cattle and human genomes by Larkin, Everts-van der Wind, Lewin, and their collaborators published in Genome Research in 2003, indicated that reuse breakpoints were more common than previously thought. Their Science paper took an even bigger crack at the random breakage theory. By comparing eight species instead of the standard two, it exposed a new, large set of reuse breakpoints.
"We're seeing more than 20 percent of breakpoints being reused," Lewin says. "That's too high to be random, and it says something." Just what is unclear at this point, though he suspects that these reuse breakpoints may be implicated in speciation, the process by which species diverge from a common ancestor.
It is clear that these reuse breakpoints are also somehow related to cancer. The team's Science research compared other researchers' work on identifying cancer-related breakpoints to the breakpoints associated with evolution. They found a correlation. "The association between evolutionary breakpoints and breakpoints in cancer implies a common underlying mechanism," Lewin says.
The final map used to make cross-species comparisons is typically a daunting undertaking. "This is totally manual," Larkin says. "You have a table to describe the structure of the human chromosome. Another describes the homologous blocks [with the other species]. Drawing the comparisons can take a month, months."
"There's been no other way of doing this other than grinding through those tables by hand or with simple [computer] scripts," Lewin says. Evolution Highway builds these maps automatically.
"This type of research used to be 80 percent moving data and getting it ready, 20 percent analysis. We're reversing those numbers," says Michael Welge, who leads NCSA's Automated Learning Group. "Let's reduce this burden, which usually falls to a grad student somewhere, and use everyone's cognitive skills on something more valuable than preparation and transformation."
Evolution Highway also offers several simple, user-oriented features that make examining the comparative maps easier. Users can look at multiple species at once, hide a given species with a click, and zoom in and out of the comparative maps, which can cover hundreds of millions of base pairs. The software's Web-services approach to delivering the tools also allows users to load the comparative analysis data in a matter of seconds and share maps easily. A custom track feature, meanwhile, allows them to incorporate new data and flag its location in an existing map.
"Here we're looking at the whole genome at once--multiple chromosomes across multiple species. The insights wouldn't have come so quickly if we couldn't throw the data at this tool from NCSA," Lewin says.
For further information: