In what can only be described as a milestone in biological and genetic engineering, scientists at Stanford University have, for the first time ever, simulated a complete bacterium. With the organism completely in virtual form, the scientists can perform any kind of modification on its genome and observe extremely quickly what kind of changes would occur in the organism. This means that in the future, current lab research that takes extremely long to perform or is hazardous in nature (dealing with lethal strains of viruses for instance), could be moved almost exclusively to a computer.
The researchers chose a pathogen called Mycoplasma genitalium as their target for modeling, out of practical reasons. For one, the bacterium is implicated in a number of urethral and vaginal infections, like its name might imply as well, however this is of little importance. The bacterium distinguishes itself by having the smallest genome of any free-living organism, with just 525 genes. In comparison, the ever popular lab pathogen, E. coli has 4288 genes.
Don’t be fooled, however. Even though this bacterium has the smallest amount of genetic data that we know of, it still required a tremendous amount of research work from behalf of the team. For one, data from more than 900 scientific papers and 1,900 experiments concerning the pathogen’s behavior, genetics, molecular interactions and so on, were incorporated in the software simulation. Then, the 525 genes were described by 28 algorithms, each governing the behaviour of a software module modelling a different biological process.
“These modules then communicated with each other after every time step, making for a unified whole that closely matched M. genitalium‘s real-world behaviour,” claims the Stanford team in a statement.
Thus, even for an organism of its size, it takes that much information to account for every interaction it will undergo in its lifespan. The simulation work was made using a 128-node computing cluster, and, even so, a single cell division takes about 10 hours to simulate, and generates half a gigabyte of data. By adding more computing power, the computing process can be shortened, however its pretty clear that for more complex organisms, much more resources might be required.
“You don’t really understand how something works until you can reproduce it yourself,” says graduate student and team member Jayodita Sanghvi.
“If you use a model to guide your experiments, you’re going to discover things faster. We’ve shown that time and time again,” said team leader and Stanford professor Markus Covert.We’d love to see this research expanded forward, which most likely will happen, but we’re still a long way from modeling a human – about 20,000 genes short.
The findings were presented in the journal Cell.
In a move that promises to bring the advantages of computer aided design (CAD) to genetic engineers, the first computer model of a complete bacterium has been produced in the US. It means researchers will soon be able to modify models of an organism's genome on a computer screen - or create artificial lifeforms - without the risks of undertaking wet biology in secure biosafety labs.
The pathogen is called Mycoplasma genitalium, a bacterium implicated in a number of urethral and vaginal infections. The bug was ripe for modelling say researchers at Stanford University in California, because it has the smallest genome of any free-living organism, with just 525 genes. By contrast, the popular lab pathogen E. coli has 4288 genes.
The modelling was undertaken by bioengineer Markus Covert and colleagues. To get the raw data for their model, they undertook an exhaustive literature review - spanning 900 research papers - to allow them to program into their model some 1900 experimentally observed behaviours and molecular interactions that M. genitalium can take part in during its life cycle.
In software terms, they found the behaviours of the 525 genes could be described by 28 algorithms, each governing the behaviour of a software module modelling a different biological process. "These modules then communicated with each other after every time step, making for a unified whole that closely matched M. genitalium's real-world behaviour," claims the Stanford team in a statement. Their research appears in the journal Cell (doi: 10.1016/j.cell.2012.05.044).
Such models will ultimately give biologists the freedom to undertake "what if" scenarios common in regular engineering - changing parameters in a genome design, say, like a civil engineer adjusts the width of a bridge deck on a computer to see what happens. As well as being experimentally useful, allowing artificial organisms and synthetic lifeforms to be created virtually (harming no-one), they could also boost biosafety by preventing accidental creations of lethal pathogens. In 2001, for instance, researchers in Australia accidentally created a lethal strain of mousepox.
In a commentary article in Cell, systems biologists Peter Freddolino and Saeed Tavazoie of Columbia University say they hope the work will soon be extended to more commonly used lab bugs like E. coli - but also warn that the technique's accuracy has yet to be demonstrated. It is unclear, they say, "how well overall behaviors will be predicted from a collection of separately obtained parameters" gleaned from hundreds of research papers.
But the US National Institutes of Health, which funded the modelling work, is excited. It believes the model a major step towards finding "new approaches for the diagnosis and treatment of disease", says James Anderson, an NIH program director.