All of the plants we eat are "genetically modified", and we did the modifying. Humans began domesticating plants in the mideast about 13,000 years ago, and since then we have continuously modified almost every plant we eat at the genetic level. The reason "genetically modified organisms" — the dreaded acronym, GMO — exists as a separate concept has to do with the way the genetic modifications are made.
The crops we consider non-GMO are modified using "classical strain improvement", which can be summed up as breeding: you select individual plants that have what you're looking for — taller stalks, larger fruits, etc. — and mate them with each other, and gradually the ideal apple or rice grain is approached. It's the same thing that's done to produce "pure-bred" dogs and horses. If you want to convince yourself that we do this, take a look at this image of corn cobs over the years, with the original Mexican maize plant at left and something more familiar at right:
According to the current definitions, although the genomes of these plants would be vastly different, at no point would any of them be considered "genetically modified organisms".
The underlying principle is that the traits we desire are produced by particular mutations or variants of genes in the genome, and by crossing plants with the right traits we enrich and even improve those variants. The process can be sped up by delibrately introducing genetic mutations using, for example, UV light or a chemical that modifies DNA. With this approach there is no way to know where in the genome the mutations will occur: they are random. Most of the mutations will kill the plant or cause undesirable changes, but a few will be beneficial, and we can select for those and breed them in the usual way.
If the language I've been using ("mutations", "genetic modification", "DNA-modifying chemicals") sounds a lot like what's used to describe GMO, that's because it is. But because the mutations are random, we cannot control where they occur, so we cannot know what changes produced the qualities we desire. Even worse, more changes will probably have been made beyond what we needed for our particular goal (e.g., redder tomatoes). We don't know what these other changes are, and it is difficult to determine if they cause any undesired changes to the plant beyond just looking by eye to see if everything seems okay.
Enter modern genetic engineering. The goal is to be faster, more precise, have more control over the process, and reduce unwanted side-effects caused by random mutations. But this is an almost entirely new field of engineering: the techniques were not worked out until the 1970s and 1980s, and still only weakly mastered. This leads to the central irony of all this, which is that what we can do using genetic engineering is a lot less drastic than what we do by classical strain improvement! As a genetic engineer, I can tell you that if you gave me the corn plant on the left and told me to turn it in to the plant on the right using rational engineering, I would have to tell you to take a hike, or just sit back and watch human civilization develop over the past ten thousand years. Seemingly simple things like fruit size (or, say, eye color for people) are often governed by numerous genes and pathways; the types of crops we have engineered have so far involved simpler modifications like inserting a single gene into corn to make it resistant to a new insecticide, or inserting three genes into rice enabling it to produce vitamin A†.
(This is not to say we will never get there. As with most technologies, genetic engineering will eventually become quite powerful, and there is of course a chance that these capabilities will be used for ill. But we are nowhere near there yet. Meanwhile, the public-policy frameworks within the field and relationships with government are being put in place so that we will be able to manage this effectively when we eventually get there.)
So, to sum up:
June 15, 2014
Let me end with an analogy: consider a malfunctioning airplane. The "classical strain improvement" approach to fixing it would be to produce several hundred copies of the machine (so far as we know, this cannot be done), then let a bunch of children loose in each one with wrenches and screwdrivers to do what they will (so far as we know, this should not be done), and finally take each plane out for a spin and go with the one that seems to work right. Never mind that we don't know what change actually fixed the problem, or how, or whether other changes were made. Contrast this with a seasoned mechanic, working from blueprints, carefully identifying the repair that needs to be made, and doing just that, with the right parts and tools. Enough said.
- We have been genetically modifying our crops for literally thousands of years
- In genetic engineering we have recently (in the last 30 years) stumbled upon a new way of doing the same thing
- This new way is more precise than the old way, and almost certainly faster and more efficient, but our capabilities are limited right now.