浓香型白酒酒醅微生物与酯类成分关联性定量模型的建立

为研究浓香型白酒出窖酒醅微生物与酯类成分的定量关系，该研究通过高通量测序分析浓香型白酒出窖酒醅优势微生物属的相对丰度，使用气相色谱-质谱联用法（GC-MS）测定酒醅中主要酯类成分的相对含量，并以优势微生物属为自变量，主要酯类物质为因变量，对建模变量进行两轮筛选，建立两者的偏最小二乘（PLS）数学模型。结果表明，经变量筛选后得到的模型包含11种酯类物质和16种优势微生物属；模型主成分个数为3时，所保留的变量均能够用选取的主成分进行解释；所得模型能够解释92.6%的自变量信息和89.4%的因变量信息，预测率为0.679，质量良好，该模型可以定量解释出窖酒醅微生物与酯类之间的关联性。

Quantitative model establishment for correlation between the microbial community and esters components in fermented grains of strong-flavor Baijiu

In order to study the quantitative relationship between dominant microorganisms and ester components in the fermented grains of strong-flavor (Nongxiangxing) Baijiu, the relative abundance of dominant microorganisms in the fermented grains of strong-flavor Baijiu in genus level was analyzed by high-throughput sequencing technology, and the relative percentage of main ester components in the fermented grains was determined by GC-MS, while the partial least squares (PLS) mathematical model was established after two rounds of model variables screening with the dominant microorganisms as independent variable and main ester components as dependent variable. The results showed that after variable filtering, the PLS model contained 11 kinds of esters and 16 kinds of dominant microbial genera. When the number of main components of the model was 3, the remained variables could be explained by these selected main components. The PLS model could explain 92.6% information of independent variables and 89.4% information of dependent variables, the predicting rate of the model was 0.679, with good quality, and the model can quantitatively explain the correlation between the microorganisms in fermented grains of unloading pits and ester components.