Rapid Assessment of the K1 Freshness Index in Refrigerated Fish by Fast Protein and Metabolites Liquid Chromatography with an Automated Peak Deconvolution Algorithm

Wider implementation of express testing methods ensuring fast, simple and cost-effective assessment of the freshness of fish and meat products is absolutely crucial for maintaining a healthy diet providing all essential nutrients for millions of people worldwide. It was reported earlier that Fast Protein and Metabolite Liquid Chromatography (FPMLC) technique implemented in a compact and affordable optoelectronic sensor is a viable solution of this problem. It detects a combined peak of adenosine triphosphate derivatives: the elution time and profile of this peak depend on the relative content of different metabolites. In the present study, an algorithm for the deconvolution of the combined metabolite peak into the individual contributions of inosine monophosphate (IMP), inosine, and hypoxanthine is presented. The deconvolution was performed using constrained least squares involving amplitude normalization, application of time masks, and estimation of the fractions of IMP, inosine, and hypoxanthine by fitting the individual peaks of the nucleotides to the combined peak. The algorithm was implemented as a Python script. Initial calibration was done on a training dataset consisting of carp muscle tissue extracts obtained at different stages of refrigerated storage; nucleotide concentration values determined by NMR spectroscopy were used as a reference. Afterwards the calibration coefficients were fixed and the algorithm was tested on an independent dataset consisting of the FPMLC chromatograms of trout samples; the K1 freshness index values were determined from the relative fractions of IMP, inosine, and hypoxanthine recovered from the FPMLC chromatograms. The results were compared with the K1 index values determined by NMR spectroscopy: relatively good agreement between the two methods were demonstrated. Thus, the application of FPMLC with a properly selected deconvolution algorithm enables rapid quality assessment of fish and meat products while presenting the result in terms of the universally accepted K1 freshness index.