Investigation of potential quality indicators for raw laver (Pyropia spp.) standardization: a collaborative approach between traditional assessment and analytical chemistry

Abstract

Introduction: Raw laver (Pyropia spp.) quality assessment is largely subjective and lacks scientific standardization.

Methods: This study established objective quality markers by integrating biochemical profiling with market-based valuation. Twenty-two samples from Seocheon, South Korea (Jan -Mar 2024) were classified into high, medium, and low quality based on auction price and total free amino acid (TFAA) content. Proximate composition and amino acid profiles were determined using AOAC methods and HPLC. Multivariate models (PCA, PLS-DA, OPLS-DA) were applied to identify key discriminants.

Results: The OPLS-DA model achieved excellent performance (R²X = 0.790, R²Y = 0.916, Q² = 0.911) and clear group separation. Alanine, glutamic acid, and aspartic acid were critical markers, with alanine showing the strongest correlation with TFAA (r = 0.93). The PLS-DA model achieved 100% classification accuracy in both training and test sets.

Discussion: These findings provide a robust scientific basis for raw laver quality grading, supporting transparent market practices and industrial standardization.

Keywords: Pyropia spp.; multivariate analysis; quality indicator; raw laver; seafood; standardization.

Figure 1

Multivariate analysis of proximate composition and amino acid profiles in raw laver (Pyropia spp.) samples collected from Seocheon between January and March 2024. (A) Principal component analysis (PCA) score plot and loading plot; (B) partial least squares discriminant analysis (PLS-DA) score plot and loading plot; (C) orthogonal partial least squares discriminant analysis (OPLS-DA) score plot and loading plot. High and low groups are indicated by different colors. R2X, R2Y, and Q2 values represent model fit and predictive ability.

Figure 2

Heatmap visualization of Pearson correlation coefficients among free amino acids in raw laver (Pyropia spp.) samples collected from Seocheon (January–March 2024), classified into three quality groups based on total free amino acid (TFAA) content: (A) high, (B) medium, and (C) low. ^*Color intensity represents the strength and direction of pairwise correlations. Positive correlations are shown in red, and negative correlations in blue.

Figure 3

Multivariate discriminant analysis of raw laver (Pyropia spp.) samples reclassified (3 groups, high; medium; low) based on total free amino acid (TFAA) content. (A) Partial least squares discriminant analysis (OPLS-DA) score plot showing separation of high, medium, and low groups (R2X = 0.700, R2Y = 0.916, Q2 = 0.911); (B) OPLS-DA score plot with TFAA vector sizing to visualize the influence of total free amino acid content on sample classification; (C) OPLS-DA biplot combining score and loading vectors, illustrating the correlation between samples [green, blue, red circles (●)] and variables (amino acids/nutritional components (●), colored circles); (D) loading contribution plot showing variable contributions for discriminating high, medium, and low TFAA groups; (E) variable importance in projection (VIP) scores ranking the most significant amino acids contributing to group classification. ^*Sample groups are color-coded: green (high), blue (medium), and red (low). Model validation parameters R2X, R2Y, and Q2 indicate model reliability and predictive capability. ^*(C) Nutrient components (e.g., glutamic acid, alanine, aspartic acid) clustered near the high-TFAA group (green) indicate their significant contribution to the group’s distinct profile, suggesting higher abundance and diversity of these compounds in high-TFAA groups samples.

Figure 4

Explained variance ratio (red) and cumulative variance ratio (blue) for the first 10 principal components (PCs) used in the PLS-DA model.

Figure 5

Confusion matrices of the PLS-DA classification model for raw laver grade prediction. (a) Confusion matrix for the training set (n = 230), (b) confusion matrix for the test set (n = 100).