Synchronous detection method for senescence quality of damaged Korla fragrant pears during storage

Area of Science:

• Agricultural Science
• Biochemistry
• Food Science

Background:

• Korla fragrant pears are prone to mechanical damage, leading to browning, senescence, and loss of commercial value.
• Understanding and predicting fruit senescence is crucial for minimizing post-harvest losses.

Purpose of the Study:

• To evaluate senescence quality in damaged pears using enzymatic and oxidative stress markers.
• To develop and optimize predictive models for damaged pear senescence during storage.

Main Methods:

• Measured superoxide dismutase (SOD), catalase (CAT), and peroxidase (POD) activities, along with superoxide anion (O2^-.) generation rate and hydrogen peroxide (H2O2) content.
• Investigated changes in these indicators in pears with varying injury levels during storage.
• Constructed and compared partial least squares regression (PLSR), support vector regression (SVR), and long short-term memory (LSTM) models for senescence prediction.

Main Results:

• All measured senescence indicators (SOD, CAT, POD, O2^-., H2O2) increased with storage time and higher damage levels.
• The Support Vector Regression (SVR) multi-output model demonstrated superior predictive performance.
• The SVR model achieved R² values above 0.95 for predicting all key senescence indicators.

Conclusions:

• The SVR model provides an effective method for synchronous detection of damaged pear senescence quality.
• Findings offer a theoretical basis for studying fruit senescence mechanisms and quality assessment.