Non-destructive prediction of banana ripeness using GC-MS-based volatile organic compounds profiling and orthogonal partial-least squares regression models.

Bananas (Musa acuminata) are among the world's most popular fruits and are climacteric, continuing to ripen after harvest. In Japan's banana supply chain, unripe green bananas are transported with ripening artificially suppressed. Upon arrival, the bananas are exposed to ethylene gas, which restarts the ripening process. Because these procedures involve artificial treatment and storage, quality control during post-harvest management is critical. However, current quality assessment methods rely on destructive testing, and real-time, non-contact monitoring technologies have not yet been established. This study hypothesized that a non-destructive model for predicting fruit ripeness could be developed using volatile organic compounds (VOCs) generated through biological reactions. VOCs emitted by bananas were collected and analyzed using gas chromatography-mass spectrometry (GC-MS). Orthogonal partial least squares (OPLS) regression analysis was performed, with VOC composition as explanatory variables and ripeness indicators (hue angle and soluble solids content, SSC) as response variables. The resulting model predicted banana ripeness indicators with high accuracy. Esters, which contribute to the characteristic sweet and fruity aroma of bananas, and their precursor alcohols were identified as key contributors to ripeness prediction. To the best of our knowledge, this study represents the first attempt to predict banana ripeness indicators (hue angle and SSC) non-destructively and non-invasively using VOCs emitted by the fruit. The developed model shows promise for application in banana quality control.