Login / Signup

Combining non-destructive devices and multivariate analysis as a tool to quantify the fatty acid profiles of linseed genotypes.

Aysel ArslanYusuf Ziya AygunMusa TurkmenNafiz CeliktasMehmet Mert
Published in: Talanta (2024)
Linseed (Linum usitatissimum L.) and linseed oil, with a fatty acid profile rich in both macro and micro elements, are recognized as functional foods due to their valuable positive effects on health. Fatty acids composition (FAC) is a key indicator in assessing the quality of linseeds. The FAC of linseed is typically determined using chromatographic methods, yielding highly accurate results. However, chromatographic methods entail drawbacks such as requiring pre-chemical processes, generating chemical waste, and being both expensive and time-consuming, similar to chemical analyses. This study focused on the feasibility of colorimeter and FT-NIRS data to determine the FAC (%), protein (%) and neutral detergent fiber (NDF %) in linseed samples. By employing the PLSR analysis based on FT-NIRS, it was determined that the ratios of stearic (R 2 val  = 0.74, RMSEP = 0.09 %), oleic (R 2 val  = 0.75, RMSEP = 0.26 %), linoleic (R 2 val  = 0.85, RMSEP = 0.58 %), linolenic (R 2 val  = 0.71, RMSEP = 1.07 %), 8,11,14 eicosatrienoic (R 2 val  = 0.77, RMSEP = 0.02 %), margaric (R 2 val  = 0.71, RMSEP = 0.01 %), myristic (R 2 val  = 0.75, RMSEP = 0.02 %), and behenic (R 2 val  = 0.74, RMSEP = 1.12 %) in linseed could be successfully predicted. Furthermore, results demonstrated that the protein (R 2 val  = 0.87, RMSEP = 0.9 %) and NDF (R 2 val  = 0.90, RMSEP = 0.6 %) content in linseeds can be successfully predicted. PLSR has demonstrated that FT-NIRS has relatively higher predictive capability compared to color models.
Keyphrases
  • fatty acid
  • healthcare
  • public health
  • high resolution
  • climate change
  • mass spectrometry
  • machine learning
  • mental health
  • tandem mass spectrometry
  • deep learning
  • artificial intelligence