Detection of Lard Adulteration in Wheat Biscuits Using Chemometrics-Assisted GCMS and Random Forest

TLDR: In this article, a method using fatty acid-based approach is used to detect lard adulterated in wheat biscuit using chemometrics and machine learning-assisted GCMS.

Abstract: Lard adulteration in food products is undesirable particularly among people with certain diet preference or religious restrictions. Previous attempts on detecting lard in wheat-based biscuits using PCR-based method were inconsistent due to the minute amount of DNA present in lard and entrapment of DNA in the starch matrix. Hence, alternative method using fatty acid–based approach is necessary. The present study aimed to detect lard adulterated in wheat biscuit using chemometrics and machine learning–assisted GCMS. Oil was extracted from the laboratory-prepared wheat biscuits using Soxhlet extraction method, converted to fatty acid methyl ester and analysed using GCMS. Principal component analysis (PCA) and hierarchical cluster analysis (HCA) were able to cluster lard, wheat biscuits and lard-adulterated samples based on their fatty acid distribution. Random forest outperformed partial least squares-discriminant analysis (PLS-DA) in sample classification. Feature selection using random forest identified two fatty acids as potential biomarkers. C18:3n6 is proposed as the potential biomarker in discriminating pure wheat biscuits and lard-adulterated biscuits due to its dose-dependent composition with lard addition.