Effects of Roasting Conditions and Cultivar on the Bioactive Compound Profile and Oxidative Stability of Cold‐Pressed White Mustard ( Sinapis alba L.) Seed Oils

ABSTRACT Roasting is increasingly used as a pre‐treatment in the production of cold‐pressed vegetable oils, yet its effects on white mustard ( Sinapis alba L.) seed oils remain poorly characterized, particularly with respect to cultivar‐dependent responses. The present study evaluated the effect of roasting conditions (140°C, 160°C, and 180°C; 5, 10, and 15 min) on the bioactive compound profile (including tocopherols, plastochromanol‐8, phytosterols, canolol, and phenolic acids) and the oxidative stability of cold‐pressed oils obtained from three white mustard cultivars differing in erucic acid and glucosinolate content: the traditional high‐erucic cultivar Nakielska, the low‐erucic cultivar Bamberka, and the cultivar Warta (low‐erucic and low‐glucosinolate content). Roasting did not significantly alter fatty acid composition but markedly increased canolol and phenolic acid content in all cultivars, with p‐coumaric acid (up to 307.53 μg/100 g) and ferulic acid (up to 203.34 μg/100 g) detected exclusively in oils from roasted seeds; sinapic acid declined systematically by up to 81%, presumably due to thermal decarboxylation to canolol. The highest canolol content was recorded in Warta oil at 140°C/10 min (418.99 μg/g). Canolol declined above 160°C in Warta and Bamberka, indicating that optimal roasting conditions for white mustard are milder than those established for rapeseed. Tocopherol content remained largely stable across all treatments, while plastochromanol‐8 increased consistently with roasting intensity. Phytosterol content was stable in Nakielska but declined progressively in Bamberka and Warta at higher temperatures. Oxidative stability (OSI) increased with roasting intensity in all cultivars, reaching the highest value in Nakielska at 180°C/15 min (18.56 h). Spearman's rank correlation analysis revealed that the browning index was the variable most strongly correlated with OSI ( r = 0.937, p < 0.001), followed by tocopherols ( r = 0.526–0.757, p < 0.01). In contrast, canolol showed no significant correlation with OSI ( r = 0.288, ns). This suggests that thermally induced browning reactions and tocochromanols contribute more substantially to oxidative stability than canolol under these conditions. Principal component analysis revealed that variation in oil quality is primarily structured by two factors: genotype, which governs lipid composition and the tocochromanol profile; and roasting conditions, which govern canolol content, phenolic acid accumulation and oxidative stability. The recommended roasting conditions depend on the production objective: for maximum canolol enrichment, 140°C/10 min is most effective, particularly for Warta; for maximum oxidative stability, 180°C/15 min yields the highest OSI values, though at the cost of canolol decline in Warta and Bamberka; a balanced strategy targeting both parameters corresponds to roasting at 140°C–160°C for up to 10 min. Sensory quality and nutritional value were not assessed in the present study and should be considered before industrial implementation.