中文摘要：

基于气相色谱-质谱（GC-MS）结合正交偏最小二乘判别分析法（OPLS-DA）,建立了筛选玉米浆中影响青霉素发酵效价的潜在标志物的方法。利用GC-MS获得玉米浆中44种物质,经过标准化、Pareto预处理后,通过OPLS-DA对样本进行模式识别,根据模型的载荷图和变量重要性因子（VIP）筛选出9个潜在的标志物,即亮氨酸、5-酮脯氨酸、天冬氨酸、柠檬酸、酪氨酸、丝氨酸、赖氨酸、苏氨酸和葡萄糖。结果表明,这些物质与青霉素的初级代谢和次级代谢密切相关,可以作为筛选青霉素发酵优质原料的潜在质量指标。

英文摘要：

A gas chromatography-mass spectrometric （GC-MS） technique together with orthogonal partial least-squares discriminant analysis （OPLS-DA） for exploring potential markers of corn steep liquor （CSL） in penicillin fermentation has been developed. 44 compositions of CSL were determined by GC-MS. Furthermore, the data sets were mean-centered and pareto-scaling before OPLS-DA analysis. According to loading plot and variable important in the projection （VIP） value, nine potential markers were screened, namely leucine, 5-oxo-proline, aspartate, citrate, tyrosine, serine, threonine and glucose. The result suggested that they were related to primary and secondary metabolism in penicillin fermentation, and considered as potential quality specifications for screening excellent raw materials.