There are other reasons for this lull. Many genetic markers have only been discovered this decade, prompting Mackill to predict a large increase in molecular breeding next decade. And, he adds, while seed firms like Monsanto and Pioneer have invested heavily in molecular breeding, none of their research has been published, due to competition.
Over the past two years, Pioneer has stressed its use of molecular breeding to improve its soy varieties, most of which are also genetically modified. The base for Pioneer's soybeans is relatively simple, and a lot of natural variation lies outside the varieties typically used, said Soper, Pioneer's soybean research director.
"In the future," Soper said, "we'll be using some of these new molecular tools to fish some needles in the haystack that we can pull out."
For a century, individual breeders, scientists and firms have bred crops for their capacity to improve yield -- the amount of crop grown. Yield is a far more complex trait than Mackill's flood tolerance. It is not a matter of one or two genes -- it takes "dozens if not hundreds of genes to get what farmers perceive as yield," Soper said.
"We've done extensive modeling to find genes that have been selected over time," he said. "Since we know that plant breeders have bred for yield, we have a theory that a lot of the genes have increased in selection over time."
These genes have had tangible yield impacts, some increasing soy's production by up to a bushel. Over the last five years, Pioneer has learned much about these individual genes, and is now probing how they interact, Soper said.
"It's not about simply adding genes and stacking them," he said. Combine two genes that separately increase yield, and suddenly the improvements disappear. Add two others together, and the effect doubles. "It's complex," Soper said.
Despite this complexity, Pioneer is promising to expand its commercial molecular breeding program to corn next year -- a crop that has proved stubbornly resistant to marker-assisted improvements.
Of maize and monkeys
Corn, also known as maize, is genetically complex -- its genome, only recently sequenced, was much more difficult to piece together than the human genome. Its genes have been active over the past 5 million years, behaving selfishly and scrambling the genome, giving the crop an incredible diversity, Cornell's Buckler said.
"There is as much diversity between any two maize varieties as between chimp and man," Buckler said. "This is why breeding efforts have been so successful in maize."
Partially because of this complexity, however, the type of molecular breeding used for scuba rice has had limited success for corn. Buckler made this clear in a recent paper looking at what genes influenced the time corn took to flower, where the many genes surveyed had little impact on the trait.
"There really are no big effect [genes], at least for flowering time," Buckler said. "That has an implication of how we're going to make progress in the future. ... [It] means we can make very powerful predictions, but also means it will be harder to figure out individual genes."
Given the limited power of individual genes in corn, Buckler has established a research method called nested association mapping. His lab grows row upon row of corn in upstate New York, crossbreeding one reference strain -- the widely grown B73 -- with 25 different varieties. (It took seven years to breed the populations.) These diverse populations, combined with high-powered computation, should allow breeding predictions for a variety of incremental improvements in traits like drought tolerance, nitrogen use and aluminum tolerance.
Buckler's lab and many others have begun to use what is considered the next step in molecular breeding, called genomic selection. First pioneered by cattle scientists earlier this decade -- there is an actual field called "bovine functional genomics" -- genomic selection capitalizes on computing power and the large number of markers now available to rapidly make breeding decisions based on every gene influencing a trait, not just a few.
"[It] allows very accurate predictions even with small effects," Buckler said.
Buckler's fields have already helped identify genes that provide a threefold increase in the vitamin A provided by corn, turning ears a brilliant orange. The crop will be used by HarvestPlus in Zambia, part of its effort to develop staples that contain nutritional, and not just yield, improvements.