August 30, 1999
Genome Tools Pass Big Test in Fruit Flies
With the effort to sequence the entire human genome speeding toward
completion, some researchers are now focusing their energy on
developing the next generation of tools that can be used to extract
valuable scientific information from the unabridged human genetic
sequences.
Howard Hughes Medical Institute investigators, Allan
Spradling at the Carnegie Institution of Washington, and Gerald
Rubin at the University of California, Berkeley, and more than two
dozen colleagues have developed and used several types of tools to
analyze the genome of the fruit fly Drosophila melanogaster,
which has provided a treasure trove of information about genes and
their function.
“Our goal was not to identify a function for every gene, but to show that it was feasible to generate large numbers of P element mutations randomly throughout the Drosophila genome and to use them to find new genes. We did that.”
Gerald M. Rubin
"When the Human Genome Project started eight years ago, the
organizers had the foresight to sequence a number of other organisms to
serve as models and as interpretive guides to the much larger human
genome," said Rubin.
"The biggest problem in the post-genomic era will be determining
what each of the thousands of human genes does," said Spradling.
"Fortunately, it has become increasingly clear that many gene functions
have been conserved during evolution, and that crucial insights into
gene function can often be gained by studying the genomes of
experimentally favorable, non-human organisms."
In 1982, Spradling and Rubin discovered how to use a transposable
element, a piece of DNA that can jump from place to place in the
genome, to engineer changes in the fly's genome. By 1988, Spradling's
group found a way to turn this piece of DNA—the P
element—into a powerful tool for determining gene function. The
idea, says Spradling, was to create large numbers of mutant fruit
flies, each containing one P element inserted within a different gene.
The observable characteristics of each such mutant strain would help
indicate the function of the particular gene that contained the single
inserted P element.
Many laboratories have been using this technique to generate mutant
flies, which Rubin and Spradling have collected under the aegis of the
Berkeley Drosophila Genome Project(BDGP), a program backed by HHMI.
Integrating the mutant fly strains within the genome project has
greatly enhanced their value, for example, by determining which gene is
disrupted in each strain and by eliminating redundant or damaged
strains. BDGP makes the fly strains available to researchers around the
world, and scientists have used the flies to characterize more than 250
Drosophila genes.
In a research article published in the September, 1999 issue of the
journal Genetics, Rubin, Spradling and their colleagues present
the results of generating and characterizing mutant fly strains that
contain P element disruptions in 1,052 different genes, more than 25
percent of Drosophila's 3,600 vital genes. The researchers
examined each fly strain for obvious changes from wild-type flies,
including phenotypes such as lethality, near-lethality and other
readily apparent physical abnormalities. Thus, says Spradling, the
current collection focuses on genes that have a major effect on the
survival and appearance of the adult organism, rather than subtle
influences on its behavior and metabolism.
"Our goal was not to identify a function for every gene, but to show
that it was feasible to generate large numbers of P element mutations
randomly throughout the Drosophila genome and to use them to
find new genes," said Rubin. "We did that." More sophisticated
examinations—for behavioral deficits or biochemical
changes—should turn up functions for even more genes since
mutations in only about a third of Drosophila genes show
phenotypes we would have recognized in the current study, he added.
Since the current study showed that P elements can target a wide range
of Drosophila genes, the researchers now believe that their
method can be used to determine the function of essentially all of the
12,000 genes that are thought to reside in the Drosophila
genome. Rubin and Spradling are now planning to greatly expand the P
element approach.
Of course, the P element disruption technique is not the only tool
that researchers are developing to analyze genomes. In a second
research article that will appear in the same issue of Genetics,
Rubin, Spradling and more than two dozen colleagues at several
institutions in the United States and England, used a wide variety of
tools to probe a 2.9 million base pair region of the Drosophila
genome.
"We chose a region of the Drosophila genome that had been
characterized in substantial detail already from a genetic perspective,
partly though the use of P elements, and for which we now have a
sequence. The idea, then, was to apply all the available tools to this
wealth of genetic and molecular information in an attempt to understand
this piece of DNA—what genes are there, what they do, how are they
organized—as completely as possible," said Rubin.
"One very interesting finding for evolutionary biology came out of
this analysis," explained Rubin, "Genes with mutant phenotypes are far
more likely to have counterparts in other organisms, including humans,
than are genes with no known mutant phenotype."
Their analysis also provided a test for the tools being developed by
computational biologists. For example, researchers have obtained gene
sequence information from dozens of organisms and have drawn some
general conclusions about what those genes "look" like when buried
within the millions of consecutive As, Ts, Gs, and Cs, the four bases
that make up an organism's chromosomes.
Computational biologists have used this information to create
software programs that scan large stretches of raw sequence
data—the exact order of the four bases in DNA—and predict
where functional genes might lie within the large stretches of DNA that
have not been well studied. Members of the BDGP's informatics team
organized a workshop at a recent international meeting of computational
biologists where they compared the experimentally generated data with
the predictions generated by several different software programs.
"This kind of competition showed that some programs worked better
than others, but the important outcome is that people will be able to
take these results and improve their software," said Rubin. "We will
have a better set of computational tools as a result."
|
