Once plantlet embryos are produced in the lab, they need to be graded so that any inferior ones are eliminated from the system. Grading is currently done by human eye and is time consuming and subjective. The tissue culture project is developing artificial intelligence (AI) to grade embryos at an early stage, to improve germination success and reduce waste.
“We can select the best embryos – those most likely to germinate – according to certain physical characteristics,” says Scion data scientist Sam Davidson, who has used machine learning to train software to accurately identify the best embryos quickly and objectively. “AI is proving to be as good as a skilled technician. The next step will be to introduce some type of robotics to physically segregate the best from the rest.”
Plans for the coming year including development of more sophisticated machine learning algorithms, and getting a lot more plants out into the nursery, to show industry the changes being made and that the project is making good progress.
