The short answer is: while the initial transformation process can be described as a “shotgun” or random approach, researchers use numerous tools and techniques to eliminate all of the plants where the insertion occurred in an undesirable location or where the insertion could be disrupting or negatively impacting surrounding genes. Then researchers conduct rigorous safety tests on the remaining plants to make sure the insertion did not produce any new proteins—except for the protein that was specifically desired by the insertion.
Following is my longer answer, which I broke down to address each facet of your question:
- Shotgun vs. precise insertion: A colleague previously answered a similar question that addresses this part of your question. You can review that information here.
- Testing: Even though insertions into a crop genome through genetic engineering have a long history of demonstrated safety, researchers select, breed and fully characterize our GM products to make sure our final products do not have any unintended insertions. Through the development and assessment processes used, the intended insertion is well characterized.
The process of producing GMOs does not produce any new proteins, apart from the specific modification intended in any given GMO. All crops produced by these methods have their modifications characterized by comprehensively assessing the GM product’s exact DNA sequence. This DNA sequence analysis ensures GMOs contain the intended, completely characterized insert and that the inserted gene did not disrupt any existing genes. Characterization of DNA gives us the information to understand impacts of gene insertion on translated proteins. We have developed a method that involves sequencing the genome of new transformations that allows us to characterize the insert and the flanking DNA at the level of individual nucleotides and codons. Moreover, our breeders examine tens of thousands of plants, whether GM or non-GM hybrids, during development of new varieties, thereby screening and eliminating any unintended effects.
- “Large numbers of toxic proteins are certain to be produced”: This statement is not supported by evidence from the thousands of years of human intervention in the breeding of crop plants. For millennia, breeding involved the selection and propagation of plants that displayed superior characteristics relative to their parents. These superior characteristics were in part the product of genetic modifications to the plant genome introduced by the mobilization of endogenous elements such as transposons, random spontaneous mutagenesis and recombination. Within the last 100 years, naturally occurring mutagenesis has been supplemented by chemical and X-ray mutagenesis that introduces far larger numbers of genetic changes per generation than are observed naturally. Although breeding activities are the product of changes to the plant genome, there has been no observed increase in the production of toxic proteins linked to breeding. Likewise, since the incorporation of a transgene into a genome is comparable to the insertion of a transposable element, there is no reason to expect that this will lead to the production of toxins.
By comparison with this history, the changes made to a crop during the production of a GMO are similar in nature but very minor in magnitude. This blog makes interesting reading in this context It has a good explanation of how mutations are used in non-GM crop breeding.