Big Data and the Future of Risk Assessment
Big Data and the Future of Risk Assessment

Big Data and the Future of Risk Assessment

Could computers determine the safety of new GM crops?

Algorithms do all kinds of behind the scenes work for us on a daily basis. Your last Google search is one example of this. Computers and their applications are things we rely upon daily to make our lives more manageable, more time efficient and less stressful. Computers, programs, applications, artificial intelligence and data crunching keep the costs of technology low, saving you and I money and time. We already rely on all this technology to keep us safe while driving, flying and operating equipment. Is it time to consider these technologies to test the safety of new crops and foods? It may result in shorting the amount of time it takes to commercialize them, help keep prices low, and lead to new innovations in our food system.

Looking from up above and down below

Through the use of drones for example plant imaging is capable of providing daily images of a specific plant for an entire growing season. Rather than relying on multiple individuals or systems to observe the impacts on the environment and growth, drones can capture this. The data gathered from these images could be used to calculate the potential risk of a new plant to impact non-target species (such as fish, bees) and on biodiversity.

Similarly, below ground interactions between the plant roots and the surrounding soil microbiome can be fully observable through the use of synchrotron beamlines. Such beamlines are able to fully visualize the growth of a plant from germination to harvest. Soil sensors can also compliment this data collection. While typically it was easier to observe from above ground, now computational accessibility is allowing researchers to better understand what’s happening below ground.

The data on a risk level

To get an approval of a new crop variety, typically three years of field trial data is required. This is standard for all new crop varieties, regardless of whether they are GM varieties or conventional varieties. Research staff go out on a weekly basis and physically assess each plot, measuring the plants, counting and inspecting them. Once three years of field trial data is gathered, the data is used to inform regulators about the potential risks of the new variety. Conventional non-GM crops are approved for production in a few months, based on an assessment of the agronomic data.

The time required to get regulatory approval for new crop GM varieties ranges from 1-5 years depending on the country doing the regulating. Canada, the US and Brazil are at the low end of the scale with it comes to the time to conduct a full risk assessment and approve the variety for commercialization, allowing it to be planted anywhere. At the other end of the scale is the EU, taking up to 5 years for a risk assessment and approval of crop variety for use as animal feed. On average, the cost of this process accounts for 25% of the total cost of developing a new crop variety and getting it approved for production. It takes considerably longer in the EU to get variety approval due to the political interference in their regulatory system.

If all the data is gathered digitally, an algorithm based on policy standards can be used to assess and compare the new varieties to existing safe and commercialized plant varieties. This computation assessment could significantly reduce the time and cost of this phase of variety development. Computers and sensors are already used in shipments of fruits and vegetables to detect the presence of bacteria that are harmful to humans, such as E. coli. When the presence of bacteria is detected in a food product prior to shipment, it is then routed out of the human food supply chain. With computers and sensors already being used to improve the safety of our food, is it time to have them do the risk assessments of our food? While this day isn’t here right now, it isn’t as far off as one might think.