Abstract
BACKGROUND/OBJECTIVES: Fig (Ficus carica L.) seed oil represents an underexplored by-product with considerable nutraceutical potential. However, systematic evaluation of genotype × environment (G × E) interactions affecting its biochemical composition remains limited. This study assessed compositional variability across fig varieties, identified metabolic trade-offs, and developed rapid authentication protocols using FTIR-ATR spectroscopy to support predictive G × E models and marker-assisted selection. METHODS: Thirty-seven fig varieties were evaluated across two consecutive harvest years (2023-2024) in Morocco. Conventional biochemical analyses measured total phenolic content (TPC), total flavonoid content (TFC), DPPH and ABTS antioxidant activities, and oil yield. FTIR-ATR spectroscopy characterized spectral variations, with ANOVA assessing effects of year, variety, and G × E interactions. Principal Component Analysis (PCA) discriminated genotypes and years. RESULTS: TPC varied substantially (16.5-115.1 mg GAE/100 g oil), declining 36% from 2023 (48.7 ± 16.6 mg GAE/100 g) to 2024 (31.2 ± 16.6 mg GAE/100 g; F = 1372.84, p < 0.001), with TFC showing parallel trends (15.6 vs. 11.8 mg QCE/100 g). DPPH activity increased 34% in 2024 (58.5% vs. 43.7%), while ABTS activity decreased 18.6% from 32.34 ± 14.28% to 26.31 ± 6.10% (p < 0.001). Oil yield decreased from 26.7% to 21.2% and negatively correlated with phenolic accumulation (r = -0.49, p < 0.001). FTIR-ATR identified diagnostic peaks (e.g., 3012, 2928 cm(-1)), with significant G × E effects (p < 0.001). PCA captured 75.4-84.5% variance, discriminating genotypes and years. Stable high-value cultivars included 'Dottato Perguerolles', 'VCR 276/49', and 'Ferqouch Jmel'. CONCLUSIONS: Genotypic differences and year-to-year environmental conditions significantly influence fig seed oil composition. The observed negative correlation between oil yield and phenolic content indicates a trade-off between lipid biosynthesis and secondary metabolism. FTIR-ATR spectroscopy coupled with multivariate analysis enables reliable variety discrimination and year differentiation, supporting the development of stable cultivars for nutraceutical applications.