Estimation and classification of popping expansion capacity in popcorn breeding programs using NIR spectroscopy

摘要: Abstract One of the most important quality traits in popcorn breeding programs is popping expansion (PE) capacity kernel, which ratio volume to weight kernel. In this study, we evaluated whether near infrared spectroscopy (NIR spectroscopy) could be used as a tool routinely predict and/or discriminate genotypes on basis their PE. Three generations (F1, F2, and F2:3) were developed three planting seasons by manual cross-pollination self-pollination. A total 376 ears from F2:3 generation selected, shelled, subjected phenotypic analysis. Genetic variability was observed F2 generations, average PE value 31.5 ± 6.7 mL g−1. prediction models using partial least square (PLS) regression developed, root mean error calibration (RMSEC) 6.08 mL g−1, while coefficient determination (RC2) 0.26. The model principal component analysis with quadratic discriminant (PCA-QDA) best for discriminating kernels low (≤30 mL g−1) those high (>30 mL g−1) an accuracy 78%, sensitivity 81.2%, specificity 72.2%. Although NIR appears promising non-destructive method assessing intact narrow populations, greater larger sample sizes would help improve robustness predictive classificatory models.