Metabolomics Combined with Machine Learning LASSO Regression to Identify Differential Biomarkers between Taigu Yam and Tiegun Yam

Abstract

This study aimed to screen out the differential metabolites between Taigu yam and Tiegun yam by metabolomics approach, and determine differential markers for predicting different yam varieties through the east absolute shrinkage and selection operator (LASSO) regression method. Ultra-performance liquid chromatography-quadrupole time of flight tandem mass spectrometry (UPLC-Q-TOF-MS/MS) was employed to analyze metabolites in two types of yams. Principal component analysis (PCA) and orthogonal partial least squares discriminant analysis (OPLS-DA) were applied to identify distinct metabolites in these two yam varieties. Additionally, the LASSO regression method was used to screen out differential markers and establish a prediction model for variety identification. The results showed that a total of 206 metabolites were identified in the two yams. PCA found that Taigu yam and Tiegun yam were clearly distinguished. OPLS-DA further screened out 56 differential metabolites. LASSO regression analysis was performed based on these differential metabolites, and five differential markers were obtained including ophiogenin 3-O-beta-L-rhamnopyranosyl-beta-D-glucopyranoside, aspartic acid, epicatechin gallate, schaftoside and gallocatechin. These differential markers were used to establish a LASSO regression prediction model for identification of Taigu yam and Tiegun yam varieties. Based on metabolomics and LASSO regression methods, this study identified differential markers between Taigu yam and Tiegun yam, constructed a prediction model to identify different yam varieties, and would provide new ideas for the identification of yam.

Keywords

• metabolomics
• machine learning
• taigu yam
• tiegun yam
• differential markers