Residue Chemistry and Predictive Microbiology Research Unit, Eastern Regional Research Center, USDA Agricultural Research Service, 600 E. Mermaid Lane, Wyndmoor, PA 19038, United States.
Int J Food Microbiol. 2014 Feb 3;171:100-7. doi: 10.1016/j.ijfoodmicro.2013.11.019. Epub 2013 Nov 23.
Predictive microbiology is an area of applied research in food science that uses mathematical models to predict the changes in the population of pathogenic or spoilage microorganisms in foods exposed to complex environmental changes during processing, transportation, distribution, and storage. It finds applications in shelf-life prediction and risk assessments of foods. The objective of this research was to describe the performance of a new user-friendly comprehensive data analysis tool, the Integrated Pathogen Modeling Model (IPMP 2013), recently developed by the USDA Agricultural Research Service. This tool allows users, without detailed programming knowledge, to analyze experimental kinetic data and fit the data to known mathematical models commonly used in predictive microbiology. Data curves previously published in literature were used to test the models in IPMP 2013. The accuracies of the data analysis and models derived from IPMP 2013 were compared in parallel to commercial or open-source statistical packages, such as SAS® or R. Several models were analyzed and compared, including a three-parameter logistic model for growth curves without lag phases, reduced Huang and Baranyi models for growth curves without stationary phases, growth models for complete growth curves (Huang, Baranyi, and re-parameterized Gompertz models), survival models (linear, re-parameterized Gompertz, and Weibull models), and secondary models (Ratkowsky square-root, Huang square-root, Cardinal, and Arrhenius-type models). The comparative analysis suggests that the results from IPMP 2013 were equivalent to those obtained from SAS® or R. This work suggested that the IPMP 2013 could be used as a free alternative to SAS®, R, or other more sophisticated statistical packages for model development in predictive microbiology.
预测微生物学是食品科学中应用研究的一个领域，它使用数学模型来预测在加工、运输、配送和储存过程中，食品中致病性或腐败微生物的数量变化，这些食品暴露于复杂的环境变化中。它在预测食品保质期和评估食品风险方面有应用。本研究的目的是描述美国农业部农业研究服务局最近开发的一种新的用户友好型综合数据分析工具——综合病原体建模模型（IPMP 2013）的性能。该工具允许用户在无需详细编程知识的情况下，分析实验动力学数据并将数据拟合到预测微生物学中常用的已知数学模型。使用文献中先前发表的数据曲线来测试 IPMP 2013 中的模型。在与商业或开源统计软件包（如 SAS®或 R）平行的情况下，比较了从 IPMP 2013 得出的数据分析和模型的准确性。分析并比较了几种模型，包括无滞后期生长曲线的三参数逻辑模型、无静止期的简化 Huang 和 Baranyi 模型、完整生长曲线的生长模型（Huang、Baranyi 和重新参数化的 Gompertz 模型）、存活模型（线性、重新参数化的 Gompertz 和 Weibull 模型）和二级模型（Ratkowsky 平方根、Huang 平方根、Cardinal 和 Arrhenius 型模型）。对比分析表明，IPMP 2013 的结果与从 SAS®或 R 获得的结果相当。这项工作表明，IPMP 2013 可以作为替代 SAS®、R 或其他更复杂的统计软件包的免费选择，用于预测微生物学中的模型开发。
