Homing in on the Minimum Genome
Scientists have long wondered how many genes are necessary to support life. This knowledge could be used to construct new forms of artificial life to efficiently produce better biofuels or drugs.
Now, computer scientists are using hypothetical synthetic bacteria to screen out inessential genes as a way to home in on the “minimum genome.” The remnants, they hope, should make good candidates for synthesizing artificial organisms and reduce the number of costly experiments required to achieve that goal.
“If your hypothetical organism does not survive the simulation, the chances are high that it would not survive in reality,” says Roberto Marangoni, PhD, professor of bioinformatics at the University of Pisa and senior author of the study in the September issue of PLoS Computational Biology.
Scientists plan to build the minimum genome by culling unnecessary genes from Mycoplasma genitalium, a bacterium with one of the smallest known genomes. At just 521 genes, Mycoplasma’s genome is about one fiftieth the size of the human genome. In an earlier attempt to find an "essential" set of genetic instructions (published in 2006), J. Craig Venter, PhD, of the J. Craig Venter Institute, and his colleagues shaved that number from 521 to 382 by disrupting each gene one at a time. They excluded from their hypothetical minimum genome all the genes whose disruption did not kill the bacteria.
Eventually, if lab scientists try to build artificial life from scratch, testing potential minimum genomes would be a time-consuming and expensive trial-and-error process. Researchers will need a way to increase the chances of hitting the right set of genes on an early try, Marangoni says.
To address that problem, Marangoni and his team created a computer simulation to test the viability of theoretical bacteria with a variety of possible minimum genomes. They gathered all chemical reactions known to take place inside Mycoplasma and ran recurring simulations of all of these reactions, assuming different sets of genes were present. The team looked for genomes that, over the course of many repeated reactions, produced a life-friendly balance of the reaction products in the cell. The simulations of some virtual creatures resulted in wildly fluctuating chemical levels, or levels that bottomed out almost immediately—conditions that would not sustain actual life.
Some previously proposed minimum genomes failed this test. These creatures’ energy supplies went to zero very quickly, Marangoni says, which is probably what “killed” them.
Marangoni’s work is promising, but not the final word, says Arcady Mushegian, PhD, director of bioinformatics at the Stowers Institute for Medical Research, “There will have to be more computer estimates of genes and how they fit together in the metabolic puzzle,” he says. “And of course ultimately the actual organism should be engineered.”