When Bats and Humans Were One and the Same
Scientists have used computer analysis to read evolution backward and reconstruct a large part of the genome of an 80-million-year-old mammal. This tiny shrewlike creature was the common ancestor of humans and other living mammals as diverse as horses, bats, tigers and whales.
Actual DNA molecules cannot survive such lengths of time. Mammal fossils from this period are extremely rare. But by tracking the course of mammalian evolution, scientists can pinpoint when a common ancestor existed and what, in general terms, it was like.
By comparing the differences between the genetic material of living mammals, the researchers have now produced what they say is a highly accurate reconstruction of a section of the ancient creature's genetic sequence.
Other scientists have used similar methods to reconstruct individual genes. But the latest study, published in the December issue of Genome Research, presents a far longer sequence. A typical gene may contain a few thousand nucleotides, the fundamental units of DNA. The new sequence spans 1.1 million nucleotides.
The authors of the study hope to use the same methods to reconstruct the entire sequence of this early mammal's genome over the next few years.
These results may inspire visions of ''Jurassic Park,'' with scientists using DNA to bring long-vanished creatures back to life.
''It's a fascinating challenge, but it's far beyond our capability at this point,'' said a co-author, Dr. David Haussler, a Howard Hughes Medical Institute investigator at the University of California, Santa Cruz.
For now, Dr. Haussler and his colleagues have only the sequence of this reconstructed DNA, not the molecule itself.
The main value of the sequence will lie in uncovering some of the crucial steps that led from these early mammals to humans today.
''It's exciting to say something about what this ancestor looked like,'' said Dr. Rasmus Nielsen, a professor of evolutionary biology at the University of Copenhagen who was not involved in the study.
Dr. Haussler and his colleagues created a computer program that could use the DNA sequences of living mammals to reconstruct the sequence of their common ancestor. The process is akin to comparing medieval manuscripts, each with its own set of typographical errors, to reconstruct the original text. Instead of typographical errors, the computer program analyzed the mutations in the DNA of living mammals. By reversing those mutations, the program uncovers what the researchers believe is the original code.
Before they used the program to reconstruct the ancestral sequence, the scientists first carried out a series of statistical tests to see how confident they could be in their program. They created mock sequences of ancestral DNA that they allowed to mutate along several evolutionary branches. From those mutated sequences, the program came up with its best estimate of the ancestral DNA.
After repeated trials, the scientists found that the program had a 98.5 percent accuracy rate -- much higher than they had hoped.
''It was a pleasant and beautiful surprise,'' Dr. Haussler said.
Other researchers familiar with the study agree. ''We were all quite surprised in the great confidence you could have,'' Dr. Nielsen said.
The scientists then used the program to reconstruct the sequence of a large chunk of DNA called the CFTR locus, which has been sequenced in a wide range of mammals. The CFTR locus is named for one of the 10 genes it contains, the cystic fibrosis transmembrane conductance regulator gene. Mutations of it have been linked to cystic fibrosis.
The scientists compared the locus in 19 species of mammals, including humans, cats and hedgehogs. Previous studies on mammal DNA indicated that the common ancestor of these species lived 80 million years ago.
The analysis showed that this common ancestor had DNA much more like our own than some of its other descendants, judging by the reconstruction.
The researchers found, for example, that 25 percent of the ancestral CFTR locus has been lost or altered in humans.
These changes were minor compared with the ones for other mammals. About 55 percent of the ancestral CFTR locus has been lost or altered in rodents.
Dr. Haussler and his colleagues are analyzing CFTR loci from a total of 37 mammal species, and they are beginning to reconstruct the entire genome of the ancestral mammal. They hope to finish in four years.
The only way to know what this ancient DNA did is to bring it back to life. It might be possible to insert the ancestral version of the CFTR locus into mice to see what proteins are produced.
''This is an expensive project, but not undoable,'' Dr. Haussler said.
Such experiments would allow researchers to glean clues about the brains of these early mammals, their ability to see colors, their metabolism and other important aspects of their biology.
Dr. Haussler is less sanguine about using an entire genome to ''retrovolve'' an 80 million-year-old mammal. ''We're a long way from actually creating a living animal from a genome,'' he said.
Dr. Haussler said he was more excited at the prospect of reconstructing an entire series of extinct genomes like those of the common ancestor of all primates and the common ancestor of all apes.
''At each stage,'' he said, ''you'll see the key events that were instrumental in defining our lineage. It's the kind of thing you dream about when you think about human evolution.''
Copyright 2004The New York Times