The Hypothetical Fungus

A lot of work has gone into reconstructing an entire human being in a computer. Computer scientists put in the precise dimensions of a person’s body, factor in biomechanics, mimic facial expressions and so on. This work gets huge amounts of hype in the press, but for all the effort and all the attention, the results so far have left me pretty unimpressed. Does watching Kevin Bacon running around without his skin in Hollow Man really make all that work worthwhile?

Frankly, I’m much more impressed with work going on in places such as the Genetic Circuits Research Group at the University of California at San Diego. Led by Bernhard Palsson , the group recreates life from the ground up–simulating the interactions of hundreds or thousands of genes in a computer. In today’s issue of the Proceedings of the National Academy of Sciences, Palsson and his colleagues offered up the newest example of this approach. They reconstructed a humble fungus–brewer’s yeast, or Saccharomyces cervisiae.

Palsson’s work would have been impossible before our genomic age. But the yeast genome is not the final answer to how these organisms work. It doesn’t tell you how they feed, how they survive hard times, or even how fast they grow. To get those answers, you have to travel up from the genome to the organism itself, which is what Palsson has been doing.

As the yeast genome has been sequenced, Palsson and his colleagues have examined each gene, to understand how it functions. We typically think of genes making proteins and leave it at that, but actually many of these proteins turn around and influence the production of other proteins. In some cases, genes can’t make their proteins at all unless another protein switches them on. In other cases, proteins can shut down the pathway that allows a protein to be made. This allows organisms to be more than just biological factories, blindly cranking out proteins. Instead, they produce proteins only as needed, in response to changes in the environment for example, or in order to take the next step in their life cycle. Genes are a lot like electrical components this way, wired together into metabolic circuits.

Palsson’s team gathered information on 708 genes in yeast (about 16% of all their genes), and worked out their role in the yeast’s metabolism. Then they turned those roles into equations (along the lines of “as production of protein A goes up, production of protein B goes down.”) They then ran these equations in a computer–simulating a significant part of the yeast’s metabolism.

To see how good their reconstruction was, they grew some virtual yeast in their computer. The virtual yeast grew at the same rate as real yeast under similar conditions. The model could predict how much glucose yeast take up, how much carbon dioxide they spewed out, and how much of the energy-carrying molecule ATP they could create along the way.

The new PNAS paper is the latest stage in a project that’s now 13 years old. Palsson and his colleagues have already simulated simpler organisms, such as the bacteria E. coli. The latest wrok represents a big step up from bacteria, however, because yeast–like us–are eukaryotes, which all have far more complicated metabolisms than bacteria. Not surprisingly, then, the virtual yeast don’t behave exactly like real ones. Palsson’s team is adding more genes to the circuits to get better performance.

Reconstructing life in a computer this way potentially has a lot more to offer than Hollow Man. It may become possible to engineer new yeasts or bacteria to make new compounds or absorb pollutants by running virtual versions of them on a computer. Eventually, it may be possible to reconstruct human cells in a computer as well. That could allow scientists to study how genetically engineered human cells would behave in the body. Ultimately, it might be possible to see how groups of cells function together–something like studying the way computers linked in a network behave. Reconstructing a whole person in a computer–trillions of cells, each with 30,000 or so genes–is insanely complex, but no one can say what the limits are to this sort of work.