DISCRIMINATION OF AÇAI FRUIT (EUTERPE OLERACEA) ACCORDING TO THEIR GEOGRAPHICAL ORIGIN USING NEAR INFRARED SPECTROSCOPY AND CHEMOMETRICS TOOLS

INTRODUCTION
The açai palm (Euterpe oleracea Martius) is widely distributed in the Amazon basin, mostly in its eastern part, the estuary. In the State of Pará (North of Brazil), the açai palm is largely predominant in the floodplains (62% of the production), but is also growing in the land forests (38%). Around 124,000 families get their first financial subsistence from the selling of the açai fruits. Açai juice is an energetic drink rich in fiber and phenolic compounds. The last ones are mainly from anthocyanin and flavonol classes, which demonstrated to be efficient in the prevention of various chronic and degenerative diseases, being also anti-inflammatory and neuroprotective. To guarantee the traceability and authenticity of the açai fruit, it is necessary to develop technological tools to guide the agricultural industry in the post-harvest monitoring and quality improvement. Near Infrared (NIR) Spectroscopy has proven to be a very efficient analytical tool once it is rapid, non-destructive and therefore suitable for the post-harvest monitoring. The aim of this work is to assess if NIR spectroscopy can be adequate to discriminating the geographical origin of açai fruits in a larger perspective for a future study – traceability in the entire Amazon basin.

EXPERIMENTAL
To conduct the first step of this study, 106 samples of açai fruit were collected from different municipalities in the North of the state of Pará (Brazil), all bordering the Amazon river or its affluents. In each municipality, at least four different farmers were selected. For each sample, phenolic compounds were extracted previously to their quantification by Folin-Ciocalteu assay. Total anthocyanins were quantified by the differential-pH method. Twenty fruits of each sample were randomly chosen to perform the NIR readings using a Phazir (Polychromix, USA) equipment collecting spectra in the 1596-2396 nm range. A total of 2120 spectra were collected and treated using Unscrambler® X 10.2 CAMO software. Pretreatments were performed to minimize noise and interference. Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) have been used to study the potential of NIR to discriminate acai fruits according to their origin.

RESULTS AND DISCUSSION
The analysis of total phenolic and anthocyanin contents showed high values of standard deviation between samples of a same municipality, demonstrating the great variability in composition of açai fruit. Fruits of Ponta de Pedras, Muaná and São Sebastião da Boa Vista municipalities showed the highest (P>0.05) values in the yield of juice among the different municipalities. The repeatability of analysis by NIR showed coefficient of variation < 5% and NIR analysis allowed to discriminate the origin. 48 samples were used as calibration set and 24 to validate the method by LDA. 90%, 83% and 75% of Ponta de Pedras, Abaetetuba and Muaná fruit samples, respectively, were successfully discriminated. In the validation test, 90% of the samples were correctly classified. Therefore, this first prospective study suggests that NIR spectroscopy could be employed to discriminate the geographical origin fruits of açai fruits.