Short project description

Perennial grain crops such as Kernza (intermediate wheatgrass) and Silphium have environmental benefits, but domestication gaps remain, including small and variable seed size and inconsistent seed quality, which can impair stand establishment. Germination and early seedling growth are key determinants of establishment success and subsequent productivity. Recent studies demonstrate that micro-CT-derived 3D internal seed morphometrics (e.g., free-space fraction, internal abnormalities, size- and density-related descriptors) can be predictive of germination outcomes, enabling non-destructive seed quality phenotyping.

The student develops a reproducible micro-CT image analysis pipeline for Kernza and Silpihum seeds to (i) segment seeds (batch-scanned if relevant), (ii) extract biologically interpretable 3D seed traits, and (iii) quantify trait diversity across population. The extracted traits will be linked to germination performance and early seedling growth metrics via interpretable statistical/ML (machine learning) models, providing practical guidance for seed quality screening and breeding-relevant phenotyping.