In the most recent issue of the Proceedings of the National Academy of Sciences USA, professor Carsten Peterson and his collaborators Björn Samuelsson and Carl Troein demonstrate how this is possible. The American physician and scientist Stuart Kauffman - a pioneer in the field, who formulated and attempted to solve the problem as early as 1967 - is their co-author.
At any given time, each of the 25,000 genes in a cell may or may not be producing a protein - each gene is 'on' or 'off', to use language from the world of computers. A gene can affect other genes, turning them 'on' or 'off'. A simple case is that two genes are controlling a third gene. To activate this third gene, both the controlling genes might need to be active, or maybe only one or the other.
"In such a simple subsystem, sixteen different rules are possible in the interaction between the genes, and a large number of different solutions can emerge for the entire network," says professor Peterson. It was systems like this that Dr. Kauffman started working with; he assumed that the different solutions corresponded to different cell types. This would also explain how the DNA can be the same in all types of cells. Unfortunately, real systems are vastly more complicated. More than two genes may be involved in activating a single gene. In the case of three controlling genes, there are already 256 different rules. And in a system of 25 genes, the number of possibilities is greater than the number of atoms in the known universe...
To find those solutions that would produce stable systems, Peterson and his collaborators have primarily used literature knowledge from the foremost guinea pig of genetics: the fruit fly. More is known about the details of the genetic network here than in humans.
"In the fruit fly one can find almost 200 rules that are canalyzing, and this property is most likely general and applicable to genetic networks in other organisms," Carsten Peterson notes. "With 'canalyzing' we mean that there is a controlling gene that decides the value of the gene it activates by being either on or off. In that case, other controlling genes don't have any effect on the activated gene. It doesn't matter whether they are on or off.
With canalyzing rules, it turns out that the networks become stable regardless of the number of controlling genes, the size of the networks and the initial state of the system." It might be added that when Stuart Kauffman first started working on this problem, he was using punch cards. Now that the problem has been solved, it was not thanks to simulations on a powerful computer - it has been sufficient with observations, logical thinking and mathematical labor.
For further information: Professor Peterson is at present in San Diego. His office phone number is 1-858-534-0824. He can also be reached by e-mail: Carsten@thep.lu.se