Sequencing a zoo.

Recent comparative sequencing and analysis of 10 genes in 13 vertebrate species has found hundreds of identical and potentially functional sequences in stretches of the genome that scientists once referred to as "junk,' DNA. A recent report claims that these sequences have been conserved through hundreds of millions of years of evolution, a fact that suggests they may perform important roles and are worthy of future study.

"This kind of focused comparison of a few genes across multiple species can filter down the vast three billion letters of the human genome to a more manageable set that can be explored for function," says report coauthor Eric D. Green, who is scientific director of the Division of Intramural Research at the National Human Genome Research Institute (NHGRI).

Most genetic research has focused on exons, sequences making up about 2% of the total human genome that code for the creation of proteins. Until recently, it was believed that only exonic sequences were functional, and that the rest of the DNA was a sort of genetic detritus, consisting of useless code such as defective copies of genes, nonsensical repeats, and the remains of disabled retroviruses (potent viruses such as HIV that can insinuate their code into the DNA of their hosts). Yet there is increasing evidence that certain sequences outside exons that make up another 2-3% of the human genome also play critical roles. However, finding these nonexonic sequences has been difficult, in part because many are smaller than exons and lack the "start" and "stop" signals that mark protein-coding regions.

In the report, published in the 14 August 2003 issue of Nature, the team describes sequencing the genomic region corresponding to a section on human chromosome 7 that contains 10 genes. The 10 genes were sequenced in the human and 12 other species: chimpanzee, baboon, cat, dog, cow, pig, rat, mouse, chicken, zebrafish, and two species of pufferfish. The sequences were then compared using two different statistical methods in a hunt for "multispecies conserved sequences," or MCSs.

A total of 1,194 MCSs were identified between the two techniques. Of these, the vast majority were nonexonic--only 244 overlapped exons. A little over half (648) were found in introns, sequences that are transcribed to messenger RNA but removed before the RNA is transcribed into proteins. The rest of the MCSs (302) were found in areas between genes. The nonexonic sequences may possibly regulate protein transcription or perform other functions.

Comparing two species usually isn't enough to find these conserved sequences, says Green. "For example, if you take the human genome and mouse genome, at forty percent of the places, the DNA is so similar that the sequences act like Velcro--they stick together, or align." The vast majority of these sequences are identical only because the two species had a common ancestor not all that long ago in evolutionary terms, and there hasn't been enough time for any nonfunctional sequences to diverge from each other. To find the small stretches of sequences that are critical to gene and organism function, it's necessary to compare multiple genomes, says Green.

"This work has had two immediate consequences," says Maynard V. Olson, director of the University of Washington Human Genome Center. "First, bioinformaticians are using these data extensively to fine-tune their methods for finding conserved sequences. The second immediate consequence is that this work is already guiding choices for whole-genome [sequencing] projects." One of the ideas of sequencing this whole "zoo," says Green, is to get a better idea of which genomes would be most cost-effective to sequence completely in terms of sorting out critical areas and looking for elements such as MCSs.

NHGRI researchers are continuing their analyses to determine which genome comparisons are most effective at finding conserved sequences, and how many comparisons are necessary to find the largest possible number of MCSs. For example, the team found that eliminating chimpanzee and baboon sequences from their 13-species analysis didn't reduce the number of MCSs found, but removing the nonmammals reduced the total by 17%. Their methods are described in a paper by Green and colleagues published in the December 2003 issue of Genome Research.

The Nature study may be of special interest to researchers studying cystic fibrosis, because one of the genes analyzed is mutated in people with that disease. "Having the data from so many species would certainly aid in the construction of animal models," says Christopher Penland, director of research for the Cystic Fibrosis Foundation. The results could also be useful in studies of gene therapy, in which viruses are engineered to transport potentially therapeutic genes. "You could use this research to look for regions in and around the gene that nature has deemed valuable and omit other areas to reduce the overall load to be carried by the virus, when the virus capacity is limited," says Penland.

Olson says such multispecies sequencing projects may also help develop better tools to analyze single-nucleotide polymorphisms, or SNPs, in humans--in effect, comparing numerous humans as opposed to numerous species. "A major issue in human genetics right now is to improve our ability to look at very large SNP databases and develop better quantitative models for determining which ones might affect function, as opposed to being background noise," he says.

The study also addresses a controversy in evolutionary genetics regarding the pace of genetic mutation. A theory called the "molecular clock" states that mutations occur at a steady pace across time, regardless of species. However, the NHGRI studies indicate that the genomes of rodents are mutating faster than those of primates, carnivores, or artiodactyls (a type of ungulate). The NHGRI study also confirms previous work indicating that primates are more closely related to rodents than they are to carnivores (such as cats and dogs) or to the hoofed artiodactyls (such as cows and pigs).