Feb 9, 2001 - based on untargeted lipidomics. Laura Righetti1,2 , Josep Rubert1 , Gianni Galaverna2, Hurkova Kamila1, Chiara Dall'Asta2, Jana Hajslova1, ...
A novel approach to authenticate whole and refined durum wheat (Triticum durum Desf.) based on untargeted lipidomics. Laura Righetti 1,2 , Josep Rubert 1 , Gianni Galaverna 2, Hurkova Kamila 1, Chiara Dall’Asta 2, Jana Hajslova 1, Milena Stranska-Zachariasova 1 1 Department of Food Analysis and Nutrition, University of Chemistry and Technology, Prague, Czech Republic 2 Department of Food and Drug, University of Parma, Parco Area delle Scienze 95/A, 43124 Parma, Italy
•Italian law establishes that pasta must be exclusively made by durum wheat semolina and water, and a maximum contamination of 3% from common wheat flour in durum wheat flour is allowed (DPR 187/01 of 9th February 2001). •This fraud has a huge impact both on quality and economy. On the one hand, this adulterated flour produces lower pasta quality since leads to a product with a scarce resistance to cooking. On the other hand, the price of durum wheat is about 25% higher compared to common wheat. •Nowadays, wheat flour is mainly authenticated by genomics and proteomics approaches. Therefore, useful tools for the detection of the adulteration of durum wheat flour with common wheat are highly required.
Is an untargeted metabolomic strategy able to authenticate common wheat in durum wheat flour ?
30 min at 240 strokes/min
Centrifugation 10 000 rpm for 7 min
1 mL of the extract was evaporated the residues were reconstituted in 1 ml of isopropanol/Methanol/Water (60:30:5, v/v)
1 g of grounded wheat 10 mL of a mixture methanol / dichloromethane (50:50, v/v)
BEH C18 (2.1x100 mm, 1.8 µm) Column temperature: 60 °C A: 5 mM ammonium formate in H2O:MeOH (95:5) + 0,1% formic acid B: 5 mM ammonium formate in iPrOH:MeOH:H2O (65:30:5) + 0,1% formic acid
Markers selection Fraud detection
Two independent sample sets were used as a preliminary screening and as a confirmatory study to verify the stability and consistency of the models obtained.
Common wheat (Bologna, Virgilio)
Confirmatory study Common wheat (Blasco)
•172 Triticum spp. samples •7 varieties •2 farming conditions •2 harvest years
Durum wheat (Odisseo)
Untargeted Lipidomics Univariate analysis
Sources of variability Farming conditions: conventional and organic
•52 Triticum spp. samples •2 varieties
ENDOSPERM •digalactosyl diglyceride BRAN DGDG 36:4
2. Relevant markers
AR 17:0 R2X=0.847,
Two growing location: Parma and Bologna
Harvest years (2014/2015)
Potential markers validation
1. Unsupervised models (PCA) Preliminary study
Durum wheat (Timilia, Miracolo, Senatore Cappelli)
R2X=0.844, Q2=0.708 COMMON
4. Admixture sample study In order to determine the method sensitivity limit, a set of artificial samples with known content of adulterant were constructed in duplicate. The percentage values of common wheat in the mixtures (15%, 10%, 5%, 3%)
68000 66000 Intensity (cps)
3. Receiver Operating Characteristic (ROC) Curve AUC (CI 95%) 95.7% AUC (CI 95%) 100%
AUC (CI 95%) 98% AUC (CI 95%) 96.7%
Area Under the Curve (AUC) 90-100 = excellent marker 80-90 = good 70-80 = fair 60-70 = poor 50-60 = fail
56000 54000 0
23 % Common 3
The results demonstrate that untargeted lipidomics, in conjunction with chemometric tools has potential as a screening tool for the detection of wheat fraud down to the 3% adulteration level. Acknowledgements: The study presented was financially supported by LR 28/98 – Emilia Romagna Region under the funding program 49, LR28/98, Call 2013 - “AMicoGrano, Analysis of the incidence of Mycotoxins on modern and ancient Grains grown under organic and conventional regimes The authors kindly thank Dr. Silvia Folloni, Dr. Roberto Ranieri, Dr. Antonio Rossetti from OpenFields Srl (Collecchio, PR).