For a while it seemed that hardly a week went by without some genetically modified crop being touted as the next great breakthrough in increasing harvests or ending world hunger.

Usually these modifications were developed by titans of unsustainable industrial agriculture and implemented in careless and ethically dubious ways. But that wasn't the worst of it; scientists quite simply don't understand enough about plant genetics, to say nothing of ecosystem dynamics, to make rip-mix-and-burn GM work safely.

Instead, the most promising advances in agricultural biotechnology tend to be the most subtle ones -- techniques that steer away from transgenic hacking and focus instead on increasing our understanding of plant genetics. Recent research from the University of Arizona's Bio5 Institute provides a good case in point.

In a study published in the July 20 issue of Nature, the U of A researchers looked at a phenomenon known as "paramutation," which underlies the tricky fact that, though the DNA is identical, genes in offspring sometimes function quite differently than genes in parents. http://www.nature.com/nature/journal/v442/n7100/suppinfo/nature04884.html

Vicki Chandler, the study's senior author, said of paramutation in a press release, "It's been known to exist for 50 years, but nobody understood the underlying mechanism." http://www.sciencedaily.com/releases/2006/07/060721202359.htm

In corn, a gene called b1 is traditionally thought to control whether plants have purple or green stalks. If one of the two b1 copies that a plant inherits from its parents is a variant known as B-I, the plant will be purple -- unless, it turns out, that the plant has two mutant versions of a gene called mop1, which earlier work showed was necessary for paramutation to happen. In the Nature study, the researchers describe how mop1 produces an enzyme called RNA-dependent RNA polymerase. RNA carries information from genes to protein-making cellular machinery, and had long been suspected of playing a part in paramutation.

The next research step is figuring out exactly how all these pieces work together to change gene function, and determining what other components are involved. Hereditary variability involves much more than genes and RNA alone: there's environmental influences, epigenetic change, nucleosome placement, vast interactions of multiple systems at varying scales that scientists have only started to comprehend.

But the corn paramutation research could help lead to a deeper, more nuanced understanding of plant development, and eventually to more sensible farming practices and technological -- or non-technological -- interventions. By providing a welcome, complexity-embracing approach to genetics, research like this is bringing us closer to wiser and more worldchanging biotechnology.