Location, year and tree age impact near-infrared (NIR) spectroscopy-based postharvest prediction of dry matter concentration for 58 apple accessions

Primary author: Soon Li Teh
Co-author(s): Jamie Coggins; Sarah Kostick; Kate Evans
Faculty sponsor: Kate Evans

Primary college/unit: Agricultural, Human and Natural Resource Sciences
Campus: Wenatchee


In apple breeding, development of cultivars with desirable eating quality and postharvest characteristics is of paramount importance. During each season, fruit are destructively sampled and evaluated for various fruit quality traits. This presents a challenge when young seedling trees do not bear sufficient fruit for destructive sampling. Alternatively, a non-destructive tool can enable prediction of fruit quality indices regardless of fruit count, thus increasing selection efficiency. In this study, near-infrared (NIR) spectroscopy was used as a non-destructive tool to correlate with destructively-derived measurements of dry matter concentration (DMC), a trait touted to be highly linked with fruit quality. The study was aimed at evaluating NIR prediction accuracy for DMC of 2,252 fruit from 58 diverse accessions at three orchard sites belonging to the Washington State University apple breeding program. Results showed that DMC values were generally predicted at high accuracies. In characterizing DMC predictive performance of within- versus between-years, both models were highly predictive and comparable, albeit slightly higher for the former. Further analysis of location × year effects revealed that location was a more important factor than year in influencing predictive performance. Finally, in cultivar-specific models, prediction made using fruit from more established trees as a calibration set consistently yielded higher prediction accuracy. This study provides a framework for understanding the impacts of location, year and tree age on NIR prediction accuracy of DMC in diverse apple breeding accessions. In addition, this work demonstrates the importance of assessing predictive performance using multiple statistical metrics.