Fish industry operators have to process fish that arrive at various postmortem times. Postmortem time constrains processing and impacts product quality, safety, and economic value. The objective identification of biomarkers is desirable to predict the postmortem day of aging and this requires a comprehensive longitudinal characterisation of postmortem aging. We analysed the postmortem aging process in trout over a 15-day window. Quantitative physicochemical measurements (pH, colour, texture, aw, proteolysis, and myofibrillar protein solubility) performed on the same fish over time revealed the levels of protein denaturation, solubility, and pH, among other parameters, change very little when assessed by conventional chemical methods. Histological analyses were performed on thin sections and revealed fibre ruptures after 7 days of storage on ice. Ultrastructures were observed by transmission electronic microscopy (TEM) and revealed that sarcomere disorganisation occurred more often after 7 days of storage. Label-free FTIR micro-spectroscopy combined with a SVM model accurately predicted the postmortem time. Spectra-based PC-DA models also enable the identification of biomarkers corresponding to Day 7 and Day 15 postmortem. This study provides insights on postmortem aging and raises prospects for the rapid assessment of trout's freshness status from label-free imaging.
Keywords: machine learning; postmortem; processed fish; spectroscopy; sustainable; trout.