From Genome Comparisons, UCSD Researchers Learn Lessons About Evolution and Cancer
10.10.03 -- In 1905, American astronomer Percival Lowell predicted the existence of a new planet he called Planet X. Lowell proved that this new planet existed even though no one had been able to see it in the sky. Twenty-five years later, astronomer Clyde Tombaugh stumbled on images of X photographed from the Flagstaff Observatory in Arizona. Today, that planet is known as Pluto.
While it took twenty-five years for astronomers to go from theory to confirmation of Pluto's existence, it took genome scientists barely three months in 2003 to confirm a revolutionary new view of what happens in the human genome to cause dramatic evolutionary changes. Now, bioinformaticians at the University of California, San Diego (UCSD) -- who posited that 'fragile' regions exist in the human genome that are more susceptible to gene rearrangements -- are collaborating with biologists to see if their new theory can yield potentially life-saving insights into diseases such as breast cancer, in which chromosomal rearrangements are implicated.
"It took only three months to go from theory to hard scientific evidence that there are regions of the genome that are subject to evolutionary 'earthquakes' over and over again," says Pavel Pevzner, who holds the Ronald R. Taylor Chair in computer science and engineering at UCSD's Jacobs School of Engineering, and is an academic participant in Calit². "That is representative of how quickly knowledge is advancing in bioinformatics, and how useful this research can be for medicine and other fields."
In June, Pevzner and UCSD mathematics professor Glenn Tesler predicted the existence of evolutionary 'fault zones' -- hotspots where gene rearrangements are more likely to occur and change the architecture of our genomes. Their work was based on computational analysis and comparison of the human and mouse genomes. In a paper in the journal Proceedings of the National Academy of Sciences (PNAS), Pevzner and Tesler estimated that these fault zones may be limited to approximately 400 fragile regions that account for only 5 percent of the human genome. While reaching that estimate using computers, the researchers were not yet able to point to specific locations in the genome where these rearrangements are more commonplace.
The PNAS paper departed from the prevailing 'random breakage' theory of evolution that has been widely held for nearly two decades, but the theory of 'fragile breakage' quickly gained acceptance. A team led by UC Santa Cruz scientists Jim Kent and David Haussler, who are widely credited for their work in the public-sector assembly of the human genome, were the first to confirm the UCSD results. In addition, for the first time, they explicitly pinpointed the location of some of the faults in the human genome.
Kent's findings were published in the September 30 edition of PNAS, along with a commentary by two pioneers in computational biology: University of Ottawa mathematician David Sankoff, and Case Western Reserve University genetics professor Joseph Nadeau. The commentary supports the original conclusions of Pevzner and Tesler. That support is all the more notable, because Nadeau is the scientist who, in 1984, originated the random breakage theory that Pevzner and Tesler rebutted. In their article, he and Sankoff acknowledge that the random breakage theory needs to be revised along the lines spelled out by Pevzner and Tesler.
Using similar computational tools, Pevzner and his post-doctoral researcher, Ben Raphael, are working with biologists at the University of California, San Francisco (UCSF) Cancer Center to analyze chromosomal rearrangements in tumors. Their October paper in the journal Bioinformatics includes an analysis that yields the first high-resolution (albeit incomplete) picture of the genomic architecture of a complex breast cancer genome.
Human cancer cells frequently possess chromosomal aberrations (such as missing an arm of a chromosome), or rearrangements, leading to changes in genomic architecture. The breast cancer MCF7 cell line is an extreme example of such aberrations, where everything went wrong and all human chromosomes but one got rearranged, fused together, or broken, as if a tall building collapsed after an earthquake. Using the recently developed End Sequence Profiling (ESP) technique developed at UCSF Cancer Center that is cheaper and quicker than outright genome sequencing, Pevzner and colleagues analyzed human MCF7 tumor cells and derived 22 genomic rearrangements implicated in cancer, most of them previously unknown. Many of them have already been experimentally confirmed at UCSF. The UCSF team has extended this work to brain, ovarian, and prostate cancer cells, generating a ten-fold increase in the ESP data that Pevzner and Raphael are now analyzing.
"When the letters of our genomic alphabet get scrambled in a single lifetime, it can be life-threatening," says Pevzner. "But we suspect that by understanding how genomic rearrangements play out over millions of years of human evolution, we may find a correlation between these phenomena -- and possibly provide biologists with new tools to study such conditions as breast cancer at the genetic level."
As soon as reconstructions of other tumor genomes are completed, Pevzner and his colleagues will investigate whether the breakpoints implicated in cancers are correlated with the breakpoints evident in human-mouse evolution from their common ancestor 75 million years ago. And as other mammalian genomes are sequenced, Pevzner and Tesler expect to use advanced computational tools to derive further insights into human evolution and cancer.
June 2003 News Release
December 2002 News Release
UCSD Bioinformatics Laboratory
Pavel Pevzner Home Page
Glenn Tesler Home Page
Proceedings of the National Academy of Sciences
PNAS Sept. 30 Commentary by David Sankoff and Joseph Nadeau (subscription required)
PNAS Sept. 30 Article by Kent et al. (subscription required)
PNAS June 24 Article by Pavel Pevzner and Glenn Tesler (subscription required)
Doug Ramsey, (858) 822-5825, firstname.lastname@example.org