Abstract The Gram-positive, anaerobic, spore-forming bacterium, Clostridium perfringens (C. perfringens) causes a variety of diseases in humans and other animals. Spore germination is thought to be the first stage of infection by C. perfringens. AGFK, a mixture of l -asparagine, d -glucose, d -fructose, and potassium ions, is an effective nutrient germinant. The objective of this study was to investigate the effects of different AGFK concentrations (0, 50, 100, 200 mM/mL) on C. perfringens spore germination. This paper proposes a novel rapid method for the measurement of spore germinability based on microscopic hyperspectral imaging technology (HSIT). The spore germination rate (Srate), the OD600% and Ca2+-DPA% of C. perfringens were determined by chemical methods under different concentrations of AGFK. The results showed that spores have a maximum germination rate of 94.59% after 80 min with 100 mM/mL AGFK. Microscopic HSIT revealed that the spectral and spatial characteristics of spores varied during the spore germination process. Multivariate analyses (GA-siPLS and GA-PLS) and the gray symbiotic matrix (GLCM) were used to extract highly correlated spectral and spatial descriptors from the time-series data from microscopic HSIT, respectively. Single spectral, spatial signals and data fusion of spectral and spatial information were then used to predict the Srate, the OD600% and Ca2+-DPA % by GA-PLS, respectively. The result show that the Srate calibration was built by GA-PLS using data fusion variables and yielded acceptable results (Rc = 0.96, RMSEC = 0.64, Rcv = 0.93, RMSEP = 0.87, Rp = 0.94). The OD600% optimal model was built by GA-PLS using image variables and yielded acceptable results (Rc = 0.93, RMSEC = 19.36, Rcv = 0.91, RMSEP = 24.36, Rp = 0.89). For Ca2+-DPA %, the model based on the fusion of spectral and imaging data was optimal. The Ca2+-DPA % calibration yielded acceptable results (Rc = 0.95, RMSEC = 49.83, Rcv = 0.93, RMSEP = 58.98, Rp = 0.92). This work demonstrates the potential of microscopic HSIT for the non-destructive detection of spore germinability. The data fusion models also significantly improved the prediction of spore germinability. In conclusion, microscopic HSIT exhibits considerable promise for nondestructive diagnostics of spore germination.

中文翻译：

---

基于显微高光谱成像技术和化学计量学的产气荚膜梭菌孢子萌发性快速测定

摘要 革兰氏阳性、厌氧、产芽孢杆菌产气荚膜梭菌 (C. perfringens) 会引起人类和其他动物的多种疾病。孢子萌发被认为是产气荚膜梭菌感染的第一阶段。AGFK 是 l-天冬酰胺、d-葡萄糖、d-果糖和钾离子的混合物，是一种有效的营养发芽剂。本研究的目的是研究不同 AGFK 浓度（0、50、100、200 mM/mL）对产气荚膜梭菌孢子萌发的影响。本文提出了一种基于显微高光谱成像技术（HSIT）的孢子发芽率快速测量新方法。产气荚膜梭菌的孢子萌发率（Srate）、OD600%和Ca2+-DPA%在不同AGFK浓度下通过化学方法测定。结果表明，孢子在 100 mM/mL AGFK 处理 80 分钟后的最大发芽率为 94.59%。显微镜下的 HSIT 揭示了孢子萌发过程中孢子的光谱和空间特征发生了变化。多变量分析（GA-siPLS 和 GA-PLS）和灰色共生矩阵（GLCM）分别用于从微观 HSIT 的时间序列数据中提取高度相关的光谱和空间描述符。然后使用单一光谱、空间信号和光谱和空间信息的数据融合分别通过 GA-PLS 预测 Srate、OD600% 和 Ca2+-DPA%。结果表明，Srate 校准是由 GA-PLS 使用数据融合变量构建的，并产生了可接受的结果（Rc = 0.96，RMSEC = 0.64，Rcv = 0.93，RMSEP = 0.87，Rp = 0.94）。OD600% 最优模型是由 GA-PLS 使用图像变量构建的，并产生了可接受的结果（Rc = 0.93，RMSEC = 19.36，Rcv = 0.91，RMSEP = 24.36，Rp = 0.89）。对于 Ca2+-DPA %，基于光谱和成像数据融合的模型是最佳的。Ca2+-DPA % 校准产生了可接受的结果（Rc = 0.95，RMSEC = 49.83，Rcv = 0.93，RMSEP = 58.98，Rp = 0.92）。这项工作证明了微观 HSIT 在无损检测孢子萌发性方面的潜力。数据融合模型还显着提高了孢子发芽率的预测。总之，微观 HSIT 在孢子萌发的无损诊断方面表现出相当大的前景。基于光谱和成像数据融合的模型是最优的。Ca2+-DPA % 校准产生了可接受的结果（Rc = 0.95，RMSEC = 49.83，Rcv = 0.93，RMSEP = 58.98，Rp = 0.92）。这项工作证明了微观 HSIT 在无损检测孢子萌发性方面的潜力。数据融合模型还显着提高了孢子发芽率的预测。总之，微观 HSIT 在孢子萌发的无损诊断方面表现出相当大的前景。基于光谱和成像数据融合的模型是最优的。Ca2+-DPA % 校准产生了可接受的结果（Rc = 0.95，RMSEC = 49.83，Rcv = 0.93，RMSEP = 58.98，Rp = 0.92）。这项工作证明了微观 HSIT 在无损检测孢子萌发性方面的潜力。数据融合模型还显着提高了孢子发芽率的预测。总之，微观 HSIT 在孢子萌发的无损诊断方面表现出相当大的前景。数据融合模型还显着提高了孢子发芽率的预测。总之，微观 HSIT 在孢子萌发的无损诊断方面表现出相当大的前景。数据融合模型还显着提高了孢子发芽率的预测。总之，微观 HSIT 在孢子萌发的无损诊断方面表现出相当大的前景。