Development of a high-performance T-RFLP system for oral flora analysis

我们开发了一种新的方法来计算多个T-RFLP样本的数据，方法是估计T-RF组合，使用从所有样本同时产生的两种不同的荧光染料进行相关分析。这一计算是基于这样的预期，即在每次分析中，来自一个全长PCR片段的两个T-RFs的比例将几乎相同。利用该方法分析了唾液中的微生物区系，并根据口腔微生物区系的表观丰度和基于峰值频率的系数加权，利用支持向量机对口腔空气中引起口臭的挥发性硫化物甲硫醇的浓度进行了预测。应用该程序对220份唾液标本进行了口腔微生物区系分析，共鉴定出24个菌群。使用ISF加权的110个样本的细菌群表观丰度训练支持向量机预测甲硫醇的存在是否超过口腔恶臭的阈值浓度，并使用剩余的110个样本对结果进行检验。根据该程序计算的细菌组成模式，对甲硫醇浓度的预测准确率为90%。

We developed a new approach for calculating data from multiple T-RFLP samples by estimating T-RF combinations, applying a correlation analysis using two different fluorescent dyes generated simultaneously from all samples. This calculation is based on the expectation that the proportions of two T-RFs from one full-length PCR fragment would be nearly the same in each analysis. Using this method, oral microflorae in saliva were analyzed and the concentrations of methyl mercaptan, a volatile-sulfur compound that causes oral malodor, in mouth air were predicted using a support vector machine (SVM) from the apparent abundances of species in oral microflora and weighted by coefficients based on the peak frequency. Using this program, the oral microflorae in 220 human saliva samples were analyzed, and 24 bacterial groups were identified. The apparent abundances of bacterial groups weighted by the ISF in 110 samples were used in train a SVM to predict the presence of methyl mercaptan above the threshold concentration of oral malodor, and the results were tested using the remaining 110 samples. Based on the bacterial composition patterns calculated by the program, the concentration of methyl mercaptan was predicted at an accuracy of 90%.