Nondestructive determination of the astringency of pollination-variant persimmons (Diospyros kaki) using near-infrared (NIR) spectroscopy and nuclear magnetic resonance (NMR) relaxometry

Authors: Milczarek, Rebecca; Liang, Peishih; WONG, TRICIA - University Of California, Davis; AUGUSTINE, MATTHEW - University Of California, Davis; Smith, Jenny; Woods, Rachelle; Sedej, Ivana; Olsen, Carl; Vilches, Ana; Haff, Ronald - Ron; Preece, John; Breksa, Andrew

Submitted to: Postharvest Biology and Technology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 11/5/2018
Publication Date: 3/1/2019
Citation: Milczarek, R.R., Liang, P., Wong, T., Augustine, M.P., Smith, J.L., Woods, R., Sedej, I., Olsen, C.W., Vilches, A.M., Haff, R.P., Preece, J.E., Breksa, A.P. 2019. Nondestructive determination of the astringency of pollination-variant persimmons (Diospyros kaki) using near-infrared (NIR) spectroscopy and nuclear magnetic resonance (NMR) relaxometry. Postharvest Biology and Technology. 149:50-57. https://doi.org/10.1016/j.postharvbio.2018.11.006.
DOI: https://doi.org/10.1016/j.postharvbio.2018.11.006

Interpretive Summary: One type of Asian persimmon (Diospyros kaki) fruit is called “pollination-variant”. Fruits of this type will behave in one of two ways: either the flesh of the fruit will be brown in color/nonastringent in taste, or it will be orange in color/astringent in taste. The former is delicious; the latter is extremely unpalatable. Persimmon growers face the problem that both types of flesh may be present in the same persimmon fruit, and external skin color is uniformly orange, regardless of the extent of astringency. Thus, in this work, we investigated two nondestructive measurement modalities - nuclear magnetic resonance (NMR) relaxometry and near-infrared (NIR) spectroscopy - for their ability to screen pollination-variant persimmons for extent of nonastringent flesh (and thus palatability). Fruit from a diverse sample set – including 6 cultivars stored for varying lengths of time and exhibiting an array of astringency levels – was subjected to the 2 nondestructive measurements and then evaluated destructively for astringency using a sequential blotting and visual sensory analysis approach. The sensory astringency scores were then used to as the response variable to construct statistical models from each modality on its own as well as NIR and NMR combined together. These models performed well, showing a maximum misclassification rate of 33%. When the most market-relevant astringency threshold was used, NIR slightly outperformed NMR. When different astringency screening thresholds were considered, combining the two modalities did not consistently improve model performance. Thus, both modalities show promise for nondestructive characterization of the flesh of pollination-variant persimmons.

Technical Abstract: Pollination-variant persimmon (Diospyros kaki) fruit exhibits 2 distinct phenotypes based on pollination status. Seeded/brown-color/nonastringent flesh is delectable and highly desirable, while unseeded/orange-color/astringent flesh is virtually unpalatable. Growers face the problem that both types of flesh may be present in the same persimmon fruit, and external skin color is uniformly orange, regardless of the extent of astringency. Thus, in this work, the nondestructive measurement modalities of nuclear magnetic resonance (NMR) relaxometry and near-infrared (NIR) spectroscopy were investigated for their ability to screen pollination-variant persimmons for extent of nonastringent flesh. Fruit from a diverse sample set – including 6 cultivars stored for varying lengths of time and exhibiting an array of astringency levels – was subjected to the 2 nondestructive measurements and then evaluated destructively for astringency using a sequential blotting and visual sensory analysis approach. The sensory astringency scores were then used to as the response variable to construct partial least squares – discriminant analysis (PLS-DA) models from spectra from each modality on its own as well as for concatenated spectra, which combined the information from both modalities. These PLS-DA models performed well, showing a maximum misclassification rate of 33%. When the most market-relevant astringency threshold was used, NIR slightly outperformed NMR. When different astringency screening thresholds were considered, combining the two modalities did not consistently improve PLS-DA model performance. Thus, both modalities show promise for nondestructive characterization of the flesh of pollination-variant persimmons.