The current approaches for process verification during sanitation of fresh produce and other minimally processed products are limited to point measurements of sanitizer concentration at discrete locations and lack rapid biological measurements to assess effectiveness of sanitation. To address this gap, this study evaluates immobilized T7 phage on anodisc membrane as a surrogate for process verification. Fourier Transform nfrared (FTIR) spectroscopy results suggested that both chlorine and peroxyacetic acid (PAA) caused phage DNA damage and protein oxidation. The Gradient Boosting lgorithm was employed to develop predictive model for sanitizer concentration levels and Escherichia coli O157:H7 inactivation. The machine learning model predicted both the effective sanitizer concentration level and bacterial reduction with receiver operating characteristic curve (ROC) values between 0.86 and 0.93. Overall, this study identified spectral measurement of phage particles in combination with machine learning approach as an effective tool for process verification.

当前在新鲜农产品和其他最低限度加工产品的卫生过程中进行过程验证的方法仅限于在离散位置对消毒剂浓度进行点测量，并且缺乏快速的生物测量来评估卫生的有效性。为了弥补这一差距，本研究评估了阳极膜上的固定化 T7 噬菌体作为工艺验证的替代物。傅里叶变换红外 (FTIR) 光谱结果表明，氯和过氧乙酸 (PAA) 都会引起噬菌体 DNA 损伤和蛋白质氧化。采用梯度提升算法开发消毒剂浓度水平和大肠杆菌的预测模型O157:H7 失活。机器学习模型预测了有效消毒剂浓度水平和细菌减少，受试者工作特征曲线 (ROC) 值介于 0.86 和 0.93 之间。总体而言，这项研究将噬菌体粒子的光谱测量与机器学习方法相结合，确定为过程验证的有效工具。