Data Analytics for profiling of food components and food authenticity analysis

Data Analytics for profiling of food components and food authenticity analysis
December 13, 2019 No Comments Featured,Slider,Webinars AOAC India

Speaker : Abhishek Mandal

About Speaker

Abhishek Mandal – Spl: Agricultural Chemicals; Detection of Pesticide Residues; Synthesis of Nano-particles; Big Data Analysis; Dsg: Scientist, ICAR-Indian Agricultural Research Institute, New Delhi.

Food authenticity is about misrepresentation by either mislabeling or by adulterating usually with lower cost material.

Typical examples are:

  • Wine authenticity –illegal sugar addition to grape juice
  • Addition of sugar to honey
  • Addition of hazelnut oil to olive oils
  • Mislabeling geographical origin (wine, olive oil)
  • Mislabeling organic food products
  • Adulteration of meat with cheaper species

Addressing Food Authentication Challenges

  • Chemically identical foods or identical chemical entities
  • Unique marker compounds rarely found -more often small analytical differences (isotopic patterns)
  • Large natural variability based on climatic conditions, fertilizers used, variety, processing.
  • Techniques must be able to distinguish small differences
  • Databases of authentic foods must be available to understand natural variability

Analytical approaches: fundamental classifications

Irrespective of the analytical technique, there are some fundamental ways of classifying the analytical approach. Many of the techniques can be used for multiple approaches, depending on how they are configured. Typical ways of classifying the approach are:

  • Targeted vs untargeted analysis
  • Specific analyte(s) vs multi-variate analysis (mva), or
  • Laboratory vs point-of-use testing.

Data Analysis: Chemometric Approach

  • Typically large number of variables involved, ie delta value, elemental composition, mass spectra, compound concentration etc
  • Multivariate analysis must be applied
  • Supervised or unsupervised methods

Unsupervised: classification of sample without any knowledge about the origin

Supervised: similarity of unknown to authentic material 

Statistical and Machine Learning Data Analysis used in food authenticity data analyses. They include:

  1. Univariate analysis methods:

ˆ Fold Change Analysis

ˆ T-tests

ˆ Volcano Plot

ˆ One-way ANOVA and post-hoc analysis

ˆ Correlation analysis

  1. Multivariate analysis methods:

ˆ Principal Component Analysis (PCA)

ˆ Partial Least Squares – Discriminant Analysis (PLS-DA)

ˆ Orthogonal-Orthogonal Projections to Latent Structures Discriminant Analysis (OPLS-DA)

  1. Robust Feature Selection Methods in microarray studies

ˆ Significance Analysis of Microarray (SAM)

ˆ Empirical Bayesian Analysis of Microarray (EBAM)

  1. Clustering Analysis

ˆ Hierarchical Clustering

  • Dendrogram
  • Heatmap

ˆ Partitional Clustering

  • K-means Clustering
  • Self-Organizing Map (SOM)
  1. Supervised Classification and Feature Selection methods

ˆ Random Forest

ˆ Support Vector Machine (SVM)

Applications/Case studies of the Statistical and Machine Learning Data Analysis used in food authenticity data analyses

Individual case studies outlining the importance and application of the various multivariate analytical tools/ chemometrics in food authenticity/food fraud detection data analysis.

About The Author

Leave a reply

Your email address will not be published. Required fields are marked *