IT IS not quite Jurassic Park, but it could prove just as interesting. This week, researchers announced they have reconstructed a million-letter-long DNA sequence from a mammal that lived 75 million years ago and was the ancestor to almost all the mammals now roaming the Earth, including people. The ultimate aim is to reconstruct the entire genome of this long-extinct species.
鈥淭his is proof of principle,鈥� says bioinformatician David Haussler of the University of California, Santa Cruz, who led the research. A reconstruction would allow biologists to trace the evolutionary changes that led to mammals as different as moose, mice and humans. And if scientists know the genomes of animals past, they will be better able to identify crucial parts of the genomes of animals present. They will also be able to synthesise ancient genes and insert them into living animals to see what function they performed.
Jurassic Park is an implausible dream, as we are unlikely to find DNA that has been preserved intact for millions of years. But Haussler has come up with an alternative route to discovering the genetic blueprint of long-extinct animals: reconstruct it by comparing the genomes of creatures still living today. 鈥淭his is probably the best we鈥檙e going to get,鈥� says Tim Hubbard, head of the human genome analysis division at the Sanger Institute near Cambridge, UK.
Advertisement
To do this, Haussler begins by lining up the genomes of different living mammals and looking for locations where all the genomes contain the same base, or DNA 鈥渓etter鈥�. Where that happens, the same base was almost certainly present in the same location in the genome of the common ancestor from which all mammals evolved. Where the genomes have different bases, working backwards down the evolutionary tree should reveal when the changes occurred, allowing the most likely ancestral state to be worked out.
To find out how accurate a reconstruction of the lost genome might be, Haussler and his team invented a DNA sequence about 50,000 bases long. They then used a computer program to simulate how, if this stretch of DNA had come from the ancestral mammal, it would have changed over 75 million years as the ancestor evolved into 20 species. The simulation used data gleaned from the human and mouse genomes on how bases become inserted, deleted or substituted over time, as well as other phylogenetic studies.
The researchers then tested their reconstruction program on this artificial data set, asking it to rebuild the ancestral genome from the 20 descendants. When compared with the original, the reconstruction proved more than 98 per cent accurate. Moreover, the errors tended to cluster around transposons, bits of 鈥渟elfish鈥� DNA that serve no apparent function. The rest was 99 per cent accurate.
With their approach validated, Haussler and his team turned to real genomes. They worked back from a stretch of DNA that has been extensively mapped in 19 mammal species, as it includes the gene CFTR, which is implicated in cystic fibrosis in people. From this sequence of nearly 2 million bases, the team reconstructed a sequence of more than a million bases that belonged to the 鈥淏oreoeutherian ancestor鈥�, the ancestor of all placental mammals except elephants, anteaters, shrews and their relatives (Genome Research, vol 14, p 2412). They found that in the genomes of primates, including humans, the sequence has changed relatively little from that of the ancestral mammal, while in rodents and cows it has evolved much more (see Graphic).
鈥淚f we can reconstruct the ancestral genome we can put a direction to evolutionary events鈥�
鈥淚鈥檓 thrilled,鈥� Haussler says. 鈥淚t鈥檚 an archaeological expedition into the DNA that captivates me.鈥� He cannot yet reconstruct the entire genome because only a handful of mammals have had their genome sequenced 鈥� far too few for his needs. Moreover, he has not yet tested whether his technique can cope with the chromosomal rearrangements that occasionally shuffle large chunks of DNA in the course of evolution, though both he and other experts suspect it can.
For biologists trying to understand the evolution of mammals, knowing the ancestral DNA sequence for even a few genes should prove immensely valuable. 鈥淚f we can reconstruct the ancestral genome, then we can put a direction to [evolutionary] events,鈥� says Manolis Dermitzakis, an evolutionary genomicist at the Sanger Institute.
The ancestral genome also has practical applications, because it reveals regions that have not evolved. Often that is because the region does something crucial, and this can indicate which are the most important parts of genes. It is also the best way to identify important regulatory regions between genes, especially if they lie far from the gene they control.
Ultimately, geneticists might be able to synthesise the recovered DNA sequences and so reconstruct actual genes. Trying to use this DNA to resurrect an organism would raise thorny technical problems, because how DNA is packed affects its function. 鈥淚 can鈥檛 say I鈥檓 optimistic we鈥檒l ever be able to do that,鈥� Haussler says. But once it becomes cheap and easy to synthesise a gene鈥檚 worth of DNA, it should be possible to recreate a few genes from our 75-million-year-old ancestor. By putting them into a mouse, for example, we could study how they work. 鈥淚t would be a fascinating experiment,鈥� Haussler says.