Abstract: This study aimed to explore the physicochemical properties and phenolic composition of the ripe berries of five color blending grape varieties, ‘Yan 73’, ‘Yan 74’, ‘Kolor’, ‘Tintorera’, and ‘Dornfelder’, and wines made from each variety by small-scale fermentation. The results showed that ‘Kolor’ and ‘Yan 73’ berries had higher contents of total anthocyanins than the other varieties, both exceeding 4 000 mg/kg fresh sample. ‘Tintorera’ and ‘Yan 73’ grapes had lower flavonol contents, which were 43.99 and 59.6 mg/kg fresh sample, respectively. The berry skin of ‘Dornfelder’ had the lowest flavanol content. Among the five varieties, the wine made from ‘Yan 73’ had the highest concentration of anthocyanins (1 827.87 mg/L), while the wine made from ‘Tintorera’ had the lowest concentrations of flavonols (19.87 mg/L), flavanols (35.34 mg/L), and phenolic acids (43.77 mg/L). Principal component analysis performed on anthocyanin contents in berries could distinguish most varieties. Furthermore, a model that was able to distinguish the berry skin and pulp of each variety well was established using orthogonal partial least squares-discriminant analysis (OPLS-DA). Most of the differentitial substances selected were F3’5’H hydroxyl substituted anthocyanins. A Pearson correlation coefficient model was constructed for the phenolic substances common to berries and wine, which showed that 70.37% of these substances had a strong correlation. The results from this study lay a foundation for dry red wine blending for the purpose of improving the color.

Key words: blending grape; wine; phenolic compounds; correlation analysis