基于水光谱组学的细菌分类检测机理与新方法

The identification and classification of bacteria is the prerequisite of real-time inspection of food microbiological safety . We have already demonstrated in previous studies that near-infrared spectroscopy (NIRS) is a useful tool for fast and non-destructive classification of bacteria. However, the detection mechanism is ambiguous. Adopting the novel concept of Aquaphotomics, this proposal aims to ascertain the structural change pattern of water molecules due to the presence of bacterial cells in the tested system leading to the understanding of the detection mechanism. Meanwhile, a new methodology for bacterial classification can be established. The spectra of different liquid systems (sterile PBS solution, bacterial suspension in PBS and milk samples) will be studied when they are individually or simultaneously disturbed by different factors involving temperature, presence of solutes as well as bacterial cells. The interactive effects on sample spectra by different factors will be studied and the important water matrix coordinates corresponding to temperature variation and presence of solutes as well as, in particular, bacteria of different types will be allocated via spectral comparison and multivariate analysis. Moreover, chemometrical algorithms will be applied to construct and select water absorption patterns (WAPs) in light of direct computation and signal decomposition. Particularly, these typical WAPs for different types can be useful to solve the problem of random distribution of bacteria in real samples. Besides, intelligent optimization algorithms will be employed to ensure the establishment of classification models where WAPs are utilized. The complement of this project will facilitate the understanding of the mechanism of microbial detection based on NIRS and developing a novel method for bacterial classification and it also can provide new insights for the detection of bio-products.

细菌的快速无损识别与分类是实现食品微生物安全实时监管的前提。在前期证明近红外光谱用于细菌分类检测可行性的基础上，本项目以12株常见菌为例，从水光谱组学视角解析细菌引起的水分子结构改变所导致的被测体系光谱变化规律，从而探明细菌检测机理，确立细菌检测新方法。研究探讨在溶质、温度和细菌种类和浓度等干扰的单独和共同作用下不同体系（PBS溶液、PBS细菌悬浊液及牛奶样本）光谱的变化规律，明确不同干扰之间的相互影响机制；通过比较及数理统计分析定位各因素所对应的重要水基坐标；采用化学计量学方法以模态构建和信号分解的角度确定对不同细菌具有特异性的水吸收模态（WAP），并利用细菌WAP特异性解决实际样本中细菌随机分布问题；同时通过使用智能优化算法建立稳健的基于WAP的细菌分类模型。该项目利用传统方法中被忽视的水吸收波段，从水光谱组学的角度剖析细菌检测机理，建立细菌检测新方法，为其它生物信息的检测提供新思路。

食源性致病菌引起的食源性疾病给人们的生活和健康带来巨大的威胁，对食源性致病菌的监测与控制在防止食源性疾病爆发和提高食品安全方面尤为重要。传统的病原微生物检测步骤繁琐，耗时耗力，远远不能满足食品安全控制时效性的要求。因此，研究出快速、准确、成本低且适用范围广泛的食源性致病菌检测技术具有重要的意义。本项目基于光谱技术的快速、无损、可实时在线检测等特点，利用傅里叶变换近红外光谱技术、傅里叶变换中红外光谱技术以高光谱技术，通过构建传统的病原微生物检测技术与光谱表征间的联系，进一步将光谱技术与化学计量学算法，应用数据挖掘算法，建立食源性致病菌快速检测模型。结果显示：以食源性致病菌为研究对象，利用傅里叶变换近红外光谱技术建立的食源性致病菌快速分类模型，随机森林算法是非常可靠的食源性致病菌检测算法，检测准确率可高达100%，筛选出来的代表水结构特征的水基坐标可以较好的用于细菌分类检测；以液态培养基中不同浓度、不同浓度食源性致病菌为研究对象，通过对比傅里叶变换近红外光谱技术与傅里叶变换中红外光谱技术对细菌种类识别和浓度检测的能力，在明确两者均可以对液态培养基中细菌样本的类别和浓度进行快速、无损的检测的同时，发现利用傅里叶变换中红外光谱技术所建立的模型性能更为优异；以大肠杆菌、金黄色葡萄球菌和沙门氏菌三种食源性致病菌为研究对象，利用高光谱技术实现了对于不同培养时间细菌菌落分类检测以及不同通用培养基上细菌菌落分类检测的研究，确定了最佳的细菌培养时间以及最佳统一模型的同时，确定了基于高光谱技术的不同通用培养基上的细菌菌落分类检测方法，通过构建水吸收模态实现了细菌的分类检测。

内容获取失败，请点击重试