Non-destructive Determination of Soluble Solids Content of Intact Jaboticaba Fruit [Myrciaria cauliflora (Mart.) Berg, cv. Açú] in Three Maturity Stages by Means of Near Infrared Spectroscopy

Jaboticaba is a native Brazilian fruit tree which bears deep purple globose berries with whitish and juicy pulp, with slightly acid and very sweet flavour. Soluble solid contents (SSC) is probably the most common quality attribute in fruit and has direct implications towards consumer acceptance. Overall, SSC determination is time consuming and does not fit in modern grading lines. As NIR systems have been implemented to measure SSC in various fruit, the objective of this study was to evaluate the feasibility of NIR diffuse reflectance spectroscopy to quantify the SSC of intact jaboticaba fruit in three maturity stages. A total of 180 jaboticaba fruit of ‘Açú’ cultivar were harvested at three maturity stages, as such: i. green, ii. breaker, and iii. completely purple. From each maturity stage 50 fruits were used as calibration set and 10 fruits as validation set. After temperature stabilization (~25°C), the spectra were collected by using a FT-IR spectrophotometer (Spectrum 100N, PerkinElmer) in the diffuse reflectance mode over the range of 4,000-10,000 cm^-1 on the epidermal surface of two different positions of each fruit (64 scans, spectral resolution of 2 cm^-1). Soluble solids content (SSC) of individual fruit was determined using A.O.A.C. (1990, 932.12) reference method. Different treatments were applied to spectra, namely multiplicative scatter correction (MSC), standard normal variate (SNV), De-Trend, and the first derivative of Savitzky-Golay. The spectral datasets were correlated with SSC by using PLS regression algorithm. Among all treatments applied to the spectra the best PLS models were obtained using SNV transformation. The PLS analysis using full cross-validation for quantitative prediction of SSC gave very good calibration models, 9 latent variables (LVs), RMSEP = 0.99%, R² = 0.70. The performance of the PLS model slightly improved when the validation group was used as test set (8 LVs, RMSEP = 0.81%, R² = 0.82). When the full cross-validation PLS model was used to predict the SSC of the validation group the RMSEP decreased to 0.79% SSC and the R² increased to 0.83. The NIR spectroscopy can be successfully used to determine SSC of intact jaboticaba fruit as a non-destructive method and could be valid and simple tool to reduce the analytical time and cost of monitoring jaboticaba quality.