The fungus Aspergillus niger doesn't bother with sex. And geneticist Dr. Arjan de Visser knows why. Sex slows down reproduction for the fungus.
When the average fitness of fungi with the same number of mutations was compared, there appeared to be a direct correlation. The more mutations the lower the fitness. Geneticists say that in this case no epistasis takes place. In other words, the mutations have no effect on each other. But it turns out they may be wrong about this. De Visser discovered this when he mapped out the fitness of the fungus three-dimensionally: the growth rate (vertical) againts all the various combinations of mutations (the horizontal plane). This produced what is called a fitness landscape.
The landscape is full of peaks and valleys: there are mutations that do well and those that don't. Which suggests that epistasis does go on after all. Mutations that lower the fitness when they operate independently can raise it when they join forces. The landscape nicely illustrates why some combinations of mutations are a dead end for the fungus, says De Visser. 'To get from a peak of low fitness to a peak of high fitness requires at least two mutations. And natural selection does not allow for such a journey.' Between two peaks there is always a valley. And what a valley means in reality is extinction.
De Visser and his theoretical colleagues in Cologne then got the fungus to engage in virtual sex. With the help of a computer they looked at how the fungus does if it travels through the same landscape with sex. And that doesn't seem to work. Simply recombining certain peaks in the landscape just creates higher mountains. 'But the combinations required do not easily exist in one population. So sex is usually disadvantageous', concludes De Visser. 'Sex delays the raising of the fitness level. It lead to useful combinations of mutations immediately being broken off again.'
Less trodden paths
Yet the fitness landscapes do not tell the whole story. De Visser: 'Many lab experiments do show that sex has real advantages. I am convinced of the existence of those landscapes. It makes sense. But it is not the whole picture. These are of course just five mutations out of many millions of possibilities. There could of course be less-trodden paths that we have missed with our sample. It depends entirely on the topography on the landscape.'