Classification of Unknown Samples by Fatty Acids

User Benefits

- Simple workflow from LabSolutions Insight™ to multivariate analysis/machine learning software - R&D by identifying new sample groups with a list of compounds critical in separating those groups. - Predict/Classify an unknown sample into the known sample groups.

Introduction

Fatty acids, particularly those with a high degree of unsaturation and a carbon backbone of a middle chain length, are known as functional nutrients. Multivariate analysis is often used in profiling oils of food origin, but it can also be applied in profiling brans to facilitate their class characterization aside from mere physical traits. Shimadzu Corporation Japan has entered the second year of the three year collaborative research with National Agriculture and Food Research Organization (NARO) on the amount of functional nutrients (e.g. fatty acids) in foods such as brans, tea leaves and rice. As a part of the collaboration, fatty acids were quantitated in 48 bran samples. While details of the results are in non-disclosure agreement, analysis workflow will be discussed in this article. Orange Data Mining (University of Ljubljana) was employed to identify three distinct groups out of the 48 bran samples and create a list of compounds that were important in classifying those clusters.

December 10, 2020 GMT