Tasselyzer, a machine learning method to quantify maize anther exertion, based on PlantCV

Abstract

Maize anthers emerge from male-only florets, a process that involves complex genetic programming and is affected by environmental factors. Quantifying anther exertion provides a key indicator of male fertility; however, traditional manual scoring methods are often subjective and labor-intensive. To address this limitation, we developed Tasselyzer — an accessible, cost-effective, and time-saving method for quantifying maize anther exertion. This image-based program uses the PlantCV platform to provide a quantitative assessment of anther exertion by capturing regional differences within the tassel based on the distinct color of anthers. We applied this method to 22 maize lines with six genotypes, showing high precision (F₁ score > 0.8). Furthermore, we demonstrate that customizing the parameters to assay a specific line is straightforward and practical for enhancing precision in additional genotypes. Tasselyzer is a valuable resource for maize research and breeding programs, enabling automated and efficient assessments of anther exertion.