Study on the tofu quality evaluation method and the establishment of a model for suitable soybean varieties for Chinese traditional tofu processing

Various characteristics of soybeans play important roles in tofu quality. In this study, soybean chemical traits and ten parameters of tofu texture and sensory characteristics were evaluated for 48 soybean varieties to determine the relationship of soybean chemical traits with tofu texture and sensory characteristics by correlation analysis. After 6 of 48 tofu variety products were set as outliers by the Box-plot method, an evaluation model for tofu overall quality was established based on ten parameters. Cluster analysis was used to classify 42 tofu varieties into five grades of edible quality based on the tofu quality evaluation model. A predictive model for the edible quality of tofu based on the chemical traits of soybeans was constructed by stepwise regression. The average relative error between the actual value and the predicted value of the validation varieties was 3.48%. The predicted quality for 11 of the 12 tested soybean varieties showed similar results to the actual tofu quality. This study suggested that total protein, glycinin (11S), glutamate, water-soluble protein, alpha-subunit, linoleic acid and tyrosine contents could be used to predict the quality of tofu. Therefore, our study provides a guideline for generating the ideal soybean varieties for producing high-quality edible tofu.