This is the reason why, if we are to keep allowing patents on genes, we have to be extremely careful on what we allow, and what we don't.
For instance, let's take Soybeans. There have been lawsuits out there claiming that someone's Soybean field had been 100% contaminated by Monsanto seeds, so that the fact that all the plants survived Roundup plantings was completely accidental. Given how difficult it is to even cross soybeans on purpose, having an entire field of plants with perfect flowers that get contaminated through pollination is just silly.
On something like corn, however, you just can't avoid cross contamination. So if a field has corn that has genes that came straight from a GMO, it's not as if one can prove wrongdoing: It's impossible to prevent that kind of contamination.
Ultimately though, we just have to figure out if we believe that patenting GMOs is something that is good for society or not. If I make a computer program and sell it, it's not really all that difficult for me to make money out of it, if people actually want it. If I tried to sell carefully engineered soybean seeds, I would not make much money, because from a few soybeans you can make millions. In two seasons, anyone that bought the seed could just compete with me without any for the R&D cost. So if a world without some kind of protection, major seed research just have to be left to public projects that don't expect any profit. You can't even protect your seed by being the only user, because very often, the seed itself is the output. Good luck selling soybeans without giving away the DNA to make more.
And it's not like we could instead change the protections to be all about the insertion of new DNA into a plant: The entire corn catalogue of Roundup ready seeds is probably comes from less than two dozen actual insertions of foreign DNA into regular plants. The tough work is to figure out which ones are actually effective, and to cross them with high yielding regular plants. So I could just take the gene that Monsanto put on a seed and add it to my own strain by using conventional breeding techniques in about 5 years, tops.
So yeah, our options are to just stop all research, move it all to non-profits, or allow some kind of patents on seeds. You could make a case for either of the three, but we have to understand that they all have their shortcomings.