Peperomia pellucida is a species known in the Amazon as “erva-de-jabuti” that has been used in several therapeutic applications based on folk medicine. Herein, we describe the classes, subclasses, and the main compounds of the leaves, stems, and roots from P. pellucida by ultra-high performance liquid chromatography coupled to high-resolution mass spectrometry associated with molecular networks, mirror plot on the GNPS library, and machine learning. These data show compounds that were annotated for the first time in the Peperomia genus, such as 2′,4′,5′-trihydroxybutyrophenonevelutin, dehydroretrofractamide C, and retrofractamide B.

Virola surinamensis is an abundant floodplain tree, popularly known as ucuúba, that grows in the Amazon. In this study, ucuúba seed oil was obtained by supercritical fluid extraction under different operating conditions, as well as Soxhlet extraction. The operating conditions for supercritical extraction were an extraction temperature of 40, 60, or 80 °C, a pressure of 350 bar, and a CO₂ mass flow of 7.9 × 10⁻⁵ kg/s. The supercritical extraction curves were fitted to mass transfer models, and the fatty acid profiles of the extracts were determined by gas chromatography. The antimicrobial activity was assessed against Candida albicans, Staphylococcus aureus, and Escherichia coli. The highest yield obtained using supercritical CO₂ was 64.39% and the lowest was 59.21%. The phytochemical analysis showed the presence of steroids, terpenes, coumarins, and phenolic compounds. All ucuúba oil samples showed antioxidant activity. Regarding the antimicrobial activity, ucuúba oil only showed activity against S. aureus.

Hancornia speciosa Gomes is a tree native to Brazil and has therapeutic potential for several diseases. Ethnopharmacological surveys have reported that the plant is used as a hypoglycemic agent and to lose weight. This study aimed to evaluate the effects of the aqueous extract from H. speciosa latex (LxHs) in a zebrafish model of diabetes. The extract was evaluated through high-performance thin-layer chromatography (HTPLC), nuclear magnetic resonance (NMR), and Fourier-transform infrared spectroscopy (FT-IR). We then tested treatments with LxHs (500, 1000, and 1500 mg/kg) by assessing blood glucose levels in alloxan-induced diabetic animals, and metformin was used as a control. The toxicity was evaluated through histopathology of the pancreas and biochemical assessment of serum levels of AST, ALT, creatinine, and urea. The extract was also assessed for acute toxicity through several parameters in embryos and adult animals. Finally, we performed in silico analysis through the SEA server and docking using the software GOLD. The phytochemical study showed the compounds cornoside, dihydrocornoide, and 1-O-methyl-myoinositol (bornesitol). The treatment with all doses of LxHs significantly decreased alloxan-induced hyperglycemia without any significant histological or biochemical abnormalities. No significant frequency of teratogenesis was observed in the embryos exposed to the extract, and no significant behavioral changes or deaths were observed in adult animals. In silico, the results showed a potential interaction between inositol and enzymes involved in carbohydrates’ metabolism. Overall, the results show a hypoglycemic activity of the extract in vivo, with no apparent toxicity. The computational studies suggest this could be at least partially due to the presence of bornesitol, since inositols can interact with carbohydrates’ enzymes.

Brazil nut oil is highly valued in the food, cosmetic, chemical, and pharmaceutical industries, as well as other sectors of the economy. This work aims to use the Fourier transform infrared (FTIR) technique associated with partial least squares regression (PLSR) and principal component analysis (PCA) to demonstrate that these methods can be used in a prior and rapid analysis in quality control. Natural oils were extracted and stored for chemical analysis. PCA presented two groups regarding the state of degradation, subdivided into super-degraded and partially degraded groups in 99.88% of the explained variance. The applied PLS reported an acidity index (AI) prediction model with root mean square error of calibration (RMSEC) = 1.8564, root mean square error of cross-validation (REMSECV) = 4.2641, root mean square error of prediction (RMSEP) = 2.1491, R²_cal (calibration correlation coefficient) equal to 0.9679, R²_val (validation correlation coefficient) equal to 0.8474, and R²_pred (prediction correlation coefficient) equal to 0, 8468. The peroxide index (PI) prediction model showed RMSEC = 0.0005, REMSECV = 0.0016, RMSEP = 0.00079, calibration R² equal to 0.9670, cross-validation R² equal to 0.7149, and R² of prediction equal to 0.9099. The physical–chemical analyses identified that five samples fit in the food sector and the others fit in other sectors of the economy. In this way, the preliminary monitoring of the state of degradation was reported, and the prediction models of the peroxide and acidity indexes in Brazil nut oil for quality control were determined.