Untargeted GC-MS metabolomics combined with multivariate statistical analysis as an effective method for discriminating the geographical origin of shrimp paste

Abstract

Abstract Traditional fermented shrimp paste is wildly recognized for its distinctive flavor, exceptional nutritional value, and numerous health benefits. The volatile components of shrimp paste can be influenced by the production process and geographical environment, thereby reflecting its unique characteristics and quality. The objective of this study was to investigate the correlation between shrimp paste produced from different geographical origins and establish a rapid and accurate method for distinguishing them. To accomplish these objectives, the volatiles of shrimp paste from various regions were extracted using headspace solid-phase micro-extraction (HS-SPME) and subsequently analyzed by GC-MS. The data is preprocessed through MS-DIAL and then subjected to multivariate statistical analysis, wherein correlation analysis, principal component analysis (PCA), partial least square-discriminant analysis (PLS-DA), and orthogonal partial least square-discriminant analysis (OPLS-DA) were collectively employed. The findings demonstrate that OPLS-DA exhibits a favorable discriminant region in the context of discriminant analysis. The integration of untargeted GC-MS metabolomics with multivariate statistical analysis offers a rapid and efficient approach for discriminating shrimp paste originating from diverse regions, which circumvents the time-intensive process of volatile compound identification.