Alba Tres (University of Barcelona) visit RIKILT
Feed identification techniques

Short Term Scientific Mission from Dr Alba Tres, University of Barcelona to host Saskia van Ruth, Institute of Food Safety – Wageningen University Research. Visit period: from 02/11/2009 to 31/01/2010 Place: Wageningen (NL)

1. Purpose of the visit

The purpose of this visit was to assess if the fatty acid composition of feeds used for the production of organic eggs was useful to identify them. Obtaining these results was interesting in order to compare them with results obtained with the corresponding eggs. Furthermore, this visit allowed the fellow to get trained in the application of multivariate statistics to assess results obtained from a large number of variables, in order to find models that allow prediction of the organic origin of products.

2. Description of the work carried out during the visit

In order to achieve these main objectives, the fellow participated in a current study at RIKILT that dealt with the development of models to predict the identity of eggs produced in The Netherlands through different production systems: organic, barn and free range eggs. A total of 48 eggs samples produced in The Netherlands had been collected, as well as the corresponding feeds that had been provided to the hens. These 48 samples included 24 samples of organic eggs (and the 24 corresponding feeds) and 24 samples of regular eggs (12 barn and 12 free-range; and the 24 corresponding feeds). The 12 barn egg (and feed) samples included 2 special eggs that were labeled as high omega-3 eggs. The fellow entered on this ongoing project performing the fatty acid analysis of feeds. Fat from feeds was extracted with chloroform/methanol (2:1, v/v). The fatty acid methyl esters were obtained as described in Guardiola et al (1994), and were determined by gas chromatography in a Varian (Palo Alto, CA) CP-3800 model gas chromatograph, fitted with a flame-ionization detector and split-splitless injector port, set at 280 ºC and 250 ºC, respectively. The split ratio was 1:30. Chromatographic separation of FA methyl esters was performed on CP-Select CB for FAME capillary column (50 m x 0.25 mm i.d.; Varian, Palo Alto, CA). Hydrogen, at 18 p.s.i., was used as carrier gas, and the oven was programmed as follows: initial temperature, 100 °C, increased at 5 °C/min to 230 °C and held for 9 min. The sample volume injected was 1 μL. Fatty acids were identified by their retention times according to those found in a fatty acid methyl ester mixture supplied by Supelco (Supelco 37 Component FAME mix, Supelco, St. Louis, MO). Results were expressed as area normalization (in %). All feeds were analyzed in triplicate.
Statistical analysis of the obtained results was performed using Pirouette (Infometrix, Bothell, WA, USA) software. Firstly, several data preprocessing were assayed: none, mean center, autoscale and variance scale. These preprocessing treatments were performed both on the relative data, as well as after logarithmic transformation of the relative data.
Principal Component Analysis (PCA) was performed after each pre-processing step to investigate if there was any natural clustering within the data. Also, Hierarchichal Cluster Analysis (HCA) was performed in order to find an optimum dendrogram or set of clusters, by grouping together the most similar samples (unsupervised), and subsequently grouping them into progressively larger and more heterogeneous units.
As the identity of feeds was known (according to the production system they were used for), a supervised clustering technique, Partial Least Squares-Discriminant Analysis (PLS-DA), was applied. Models created were validated by cross validation: 10% of the samples was removed each time, the model was built again with the remaining samples, and it was used to predict the identity of the removed samples. This was repeated until predictions were obtained for all samples.

3. Description of the main results obtained

Multivariate statistical techniques were applied to analyze feed fatty acid composition. When PCA was conducted without data pre-processing, or when mean centering was applied to pre-process samples, 3 factor models that were built were highly influenced by the two feeds used for the production of high omega-3 eggs. Therefore, these samples were excluded from the main data set.
With the remaining data set, models were built using PCA after different sample preprocessing. All built models explained more than 95% of variance (except when data was autoscaled). Some natural clustering appeared, but it did not correspond with the different production systems for which the feeds were used. However, all feed used for organic egg production were grouped together (except one sample that was out of the main scores cloud), but this group also included some feeds used for the production of barn and free-range eggs.
HCA was performed with the data set (without omega-3 feeds). It did not group BIO19 sample with any group. The rest of data was divided into two main groups,and then into three, according to feed fatty acid composition, which corresponded with the clustering observed in the PCA.
As the production systems for which feeds were used was known, we could apply a PLS-DA to the data set (omega-3 feeds excluded). When it was applied to none preprocessed data, or with mean centered data, It predicted correctly feeds used for organic production (except BIO19, which showed a different pattern from the rest of organic feeds). However for regular feeds, predictions were not as successful.
Model with logarithmic transformed data was also built using PCA. Although it explained more than 95% of variance, and it also revealed some clustering within the data, these clusters did not match with the production system for which feeds were used. Predictors done by means of PLS-DA were even worse than those obtained with models without logarithmic transformations.

4. Conclusion

The STSM has been very successful because it has allowed to analyzed the fatty acid composition of feed belonging to an ongoing project at RIKILT. Successful models were built and cross validated, to predict the identity of feeds used for organic egg production. Furthermore, this STSM has allowed that the fellow got trained in the application of several multivariate statistic techniques, which were new for her. This will give her the possibility to apply this techniques in her home institution in future research.

5. Future perspectives

During this STSM feed fatty acid composition was obtained, which will allow to compare these data with the fatty acid composition of the corresponding eggs in order to establish possible relationships and classifications. Also, other feed and egg components will be analyzed, for instance, the carotenoid profile. Then it will be investigated if the carotenoid profile (alone or combined with the fatty acid profile) would be able to discriminate between feeds for organic and
regular egg production.
• Future collaboration with host institution
After the STSM, the collaboration with the host institution is expected to be continued later in 2010 after return of the fellow to Spain. The fellow and the host institution were granted a Marie Curie IEF fellowship, which has entered recently the last part of the negotiation process.
• Projected publications/articles resulting or to result from the STSM. It is planned to publish the obtained results at least in one peer-reviewed paper, which is planned to be sent within 1 month to Animal Feed Science and Technology, with the tentative title: “Feed influence on the fatty acid composition of organic eggs: prediction of their identity using multivariate statistics”.
Also, results will be presented at the 101st AOCS Annual Meeting and Expo (Phoenix, USA, May 2010) in form of a poster (Abstract submitted, waiting for acceptance).
• Confirmation by the host institute of the successful execution of the mission. Dr. Alba Tres familiarized herself with the laboratory very quickly. She participated in this innovative study on organic feed authentication with fingerprint profiling techniques. She accomplished a considerable amount of work in a short time. The results will be certainly worthwhile to be published in a peer-reviewed scientific journal. We are very pleased with the stay of Dr. Tres. This STSM has been a fruitful collaboration and exchange of scientific information.
·References: Guardiola, F.; Codony, R.; Rafecas, M.; Boatella, J.; Lopez, A. Fatty acid composition and nutritional value of fresh eggs, from large- and small-scale farms. J. Food Comp. Anal. 1994, 7, 171–188.

Alba Tres STSM
Design & Programming: Zenlex