"The Modern View of Domestication," a special feature of The Proceedings of the National Academy of Sciences (PNAS) published April 29, raises a number of startling questions about a transition in our deep history that most of us take for granted. At the end of the last Ice Age, people in many spots around the globe shifted from hunting animals and gathering fruits and tubers to cultivating livestock and plants.

It seems so straightforward and yet the more scientists learn, the more complex the story becomes. Recently, geneticists and archeologists working on domestication compared notes and up popped a question of timing. Did domesticating a plant typically take a few hundred or many thousands of years?

Genetic studies often indicate that domestication traits have a fairly simple genetic basis, which should facilitate their rapid evolution under selection. On the other hand, recent archeological studies of crop domestication have suggested a relatively slow spread and fixation of domestication traits.

In this special issue of PNAS, Washington University in St. Louis biologist Ken Olsen, PhD, and colleagues ask whether complex genetic interactions might have slowed the rate at which early farmers were able to shape plant characteristics, thus reconciling the genetic and archeological findings.

Olsen, associate professor in the Department of Biology in Arts and Sciences, together with colleagues from Oklahoma State University and the University of Guelph in Ontario, Canada, conclude that these interactions are not a key factor in domesticated plants. The process of domestication, Olsen said, favored gene variants (alleles) that are relatively insensitive to background effects and highly responsive to selection.

But finding these alleles in the first place must have difficult, Olsen said. Only a subset of the genes in the wild population would have reliably produced a favored trait regardless of the crop variety into which they were bred and regardless of where that crop was grown. So the early stages of domestication might have been beset by setbacks and incomprehensible failures that might help explain the lag in the archeological record.

"What we are learning suggests there's a whole lot of diversity out there in wild relatives of crop plants or even in landraces, varieties of plants and animals that are highly adapted to local conditions," Olsen said, "that wasn't tapped during the domestication process."

"These plant populations could provide the diversity for continued breeding that is going to be very important as the world faces climatic change," he said. "This is why it is important we understand the early stages of domestication." Two possible speed bumps
Many crops are distinguished from their wild ancestors with a suite of traits called the domestication syndrome. This includes seeds that remain attached to the plant for harvesting (a trait called nonshattering), reduced branching and robust growth of the central stem and bigger fruits, seeds or tubers.

Over the past 20 years, researchers have begun to identify the genes that control some of the most important domestication traits, no easy task in the days before rapid sequencing, because they had to start with plant traits and work back to unknown genes.

This work showed that many domestication traits were under the control of single genes. For example the gene teosinte branched1 (tb1) converts highly branched teosinte plants into single stalks of corn.

But the seeming importance of single genes could have been an artifact of the method used to identify domestication genes, which required the researcher to pick "candidate" genes and, perhaps, prematurely narrow the search, overlooking indirect genetic effects.

"Little is known about the underlying genetics of domestication," Olsen said. "We decided to look at genetic mechanisms for modifying plant phenotypes that hadn't been explored before, in part because not much data is available."

The new work examines the possibility that two indirect effects - the influence of the genetic background on the expression of a gene (called epistasis) and the effects of the environment on the expression of genes - might have slowed the selection of plants with the desired traits.

Epistasis and environmental effects in domestication genes

By selecting animals for coat color, animal breeders may have stabilized certain epistatic and environmental interactions in companion animals (see photos at right). But when the plant scientists looked at comparable genetic mechanisms in domesticated plants, they found the reverse to be true. Farmers seem to have selected for plant variants that were insensitive to epistatic and environmental interactions.

Shattering in domesticated foxtail millet provides an example of insensitivity to epistasis. Branching in maize illustrates insensitivity to environmental effects.

Shattering in foxtail millet and its wild ancestor, green millet, is controlled by two stretches of DNA containing or linked to genes that underlie this trait, a major one called QTL 1 and a minor one called QTL2.

In this as in other epistatic interactions, the effect of an allele at one location depends on the state of the allele at the other location. But when wild and domesticated plants are crossed, these "genetic background effects" are not symmetric.

Shattering in plants with a wild green-millet allele at the QTLI location depends on the allele at the QTL2 location. In contrast, shattering in plants with the foxtail-millet allele at QTL1 is unaffected by the allele at the QTL2 location.

In the limited number of examples at their disposal, the scientists found it to be generally true that that domesticated alleles were less sensitive to genetic background than wild alleles. The domestication genes, in other words, tended to be ones that would produce the same result even if they were introduced into a different crop variety.

Teosinte provides a good example of the sensitivity of gene expression to the environment. Teosinte is strongly affected by crowding. When a teosinte plant with a wild tb1 gene is repeatedly backcrossed with maize, it produces highly branched plants in uncrowded growing conditions but plants with smaller lateral branches when it is crowded.

Again, however, the effect is not symmetric. The domesticated trait is less sensitive to the environment than the wild trait; plants with the domesticated tb1 gene allele are unbranched whether or not they are crowded.

Unlike companion-animal breeders, early farmers seem to have selected domestication-gene alleles that are insensitive to genetic background and to the environment. This process would have been slow, unrewarding and difficult to understand, because the effects of gene variants on the plant weren't stable.

But once sensitive alleles had been replaced with robust ones, breeders would have been able to exert strong selection pressure on plant traits, shaping them much more easily than before, and the pace of domestication would have picked up.

No wonder the archeological record indicates there were false starts, failed efforts and long delays.

"The Modern View of Domestication," a special issue of PNAS edited by Greger Larson and Dolores R. Piperno, resulted from a meeting entitled "Domestication as an Evolutionary Phenomenon: Expanding the Synthesis," held April 7, 2011, that was funded and hosted by the National Evolutionary Synthesis Centre (National Science Foundation EF-0905606) in 2011.