Mutant screening is an important procedure in plant breeding. Identifying mutants that cope better with given environmental conditions, or that deliver more and better harvest is essential to develop new varieties. Challenges are to cope with climate change conditions or to adapt to sustainable cropping systems.

High throughput phenotyping technologies can contribute to mutant screening as they automatically analyze broad ranges of phenotypic traits. However, discussions are ongoing whether automatic phenotyping does the rating as reliable and accurate as experienced breeders do. As consequence, traditional methods still prevail despite the availability of digital technology.

A research group compared screening by digital phenotyping to human mutant screening using a radiation-induced Arabidopsis mutant population as model. They screened for phenotypic traits using a LemnaTec phenotyping system for image acquisition and image analysis. Visible light images and fluorescence images were taken and analyzed for plant morphology, color, and fluorescence intensity. These phenotypic fingerprints were applied to identify mutants.

In conclusion, the automatic phenotyping system turned out to be more accurate than human rating, pointing out that applying such techniques is valuable in screening procedures. However, researchers found that in the given model study the amount of false positives was higher for the automatic phenotype screening compared to the human rating. Advanced image processing algorithms, particularly combined with machine learning might overcome this issue. The researchers recommend that in mutant screenings automatic phenotyping techniques should be applied to as a first line of selection to narrow down the amount of plants that need to undergo human rating as second step.