Scientists have urged caution over a study which may have found a so-called "superweed" growing at a site where GM crops had been trialled.
Cultivated oilseed rape has a number of wild relatives
The charlock, a relative of oilseed rape, failed to shrivel up when daubed with the herbicide used to manage a biotech crop grown in the same field.
The creation of wild plants that pick up the traits of engineered crops has long been feared by anti-GM groups.
But researchers said their work showed the chances of such transfer were slim.
What is more, they argued, the study reinforced the view that the environmental impact was negligible.
"Herbicide-tolerant weeds tend to under-perform compared with wild type, so unless all its competitors have been sprayed out with the same herbicide, it won't thrive," commented Dr Les Firbank, who led the consortium of scientists on the recent UK Farm-Scale Evaluations (FSEs) of genetically modified plants.
"There's lots of evidence for that," he told the BBC News website.
The study was conducted by Centre for Ecology and Hydrology (CEH) researchers.
It looked for any evidence that a genetic trait in an oilseed rape, engineered to be resistant to a particular herbicide called Liberty, would pass to near-relatives growing wild in the field or at the margin.
The degree to which such transfer is possible informs the debate about superweeds, which some have claimed could upset ecological relationships in the countryside and so harm biodiversity.
The CEH team collected more than 95,000 seeds of wild relatives in and around the FSE trial sites and grew them up in greenhouses. These plants were then sprayed with Liberty (a glufosinate ammonium) to see if they had acquired herbicide tolerance - through their parents being pollinated by the GM rape.
The scientists found just two plants, of Brassica rapa or turnip rape, that showed resistance to the treatment; a rate of 0.000021.
But Brassica rapa is a very close relative of farmed oilseed rape and the discovery of some gene flow is not a huge surprise, say the scientists.
The CEH team also toured fields, daubing Liberty on the tissues of weeds and looking for the expected signs of die-back.
The researchers found just one weed - what they believe was a charlock (Sinapis Arvensis) - which showed no reaction to the application.
DNA analysis on a leaf sample confirmed the gene trait from the engineered oilseed rape was present, but when the researchers returned the following year to the same field they could find no herbicide tolerance in seedlings of the charlocks growing there.
Nonetheless, anti-GM group Friends of the Earth believes the existence of just one tolerant charlock should merit major concern.
It said that if GM oilseed rape were grown commercially, herbicide-resistant weeds could become widespread.
FoE argued that farmers would then have to use more - and more damaging - weedkillers to get rid of them, with knock-on impacts on the environment.
"The government's trials have already shown that growing GM crops can harm wildlife. Now we're seeing the real possibility of GM superweeds being created, with serious consequences for farmers and the environment," commented FoE's GM campaigner Emily Diamand.
The FSEs were the largest study of their kind ever undertaken
Environment Minister Elliot Morley said: "Even if a hybrid did once exist, it has disappeared. We do however need to improve our understanding of all aspects of gene transfer and this means we must take this into account with individual GM applications.
"Our top priority is to safeguard human health and the environment. There are no trials of GM oilseed rape in the UK at the moment. No consents for commercial cultivation in the EU have been issued and there are none in the pipeline."
The £6m FSEs were described as the biggest ecological experiment in the world and a model for measuring the impact of new farming techniques on the environment.
The results for four types of engineered crops - a spring-sown oilseed rape, a winter-sown oilseed rape, a sugar beet and a maize - were tested over a period of three years.
All were engineered to be resistant to a particular herbicide, which meant they would continue to prosper when the weedkiller was applied to the "pest" plants in the field.
Only the maize came through the trials with approval because the field management used to cultivate the bitotech crop appeared to be kinder on wildlife than the regime employed on the conventional maize grown as a controlled comparison.