German scientists developed a determinant of organic chicken eggs using AI

A team of researchers from the German Institute of Food Technology (DIL) in Quakenbrück has developed an innovative method by which measurements can be used to determine whether an egg actually comes from organically farmed laying hens. To do this, the researchers examined about 4,500 egg samples, writes the German poultry portal Gefluegelnews.

AI-powered reference database

To prove the content type, the researchers analyzed egg yolk extract samples obtained from eggs using 1H NMR spectroscopy. The spectra obtained in this way represent a very specific pattern of the egg sample , a fingerprint with a lot of information. 

Using multivariate data analysis, machine learning and artificial intelligence , the researchers identified characteristic patterns for each type of livestock production and created a database and authenticity model based on these reference spectra. By comparing the 1H-NMR spectrum of unknown egg yolk samples with a model from reference spectra, they were able to demonstrate the actual behavior of laying hens.

The model calculated to classify eggs from conventional and organic farming achieved an accuracy of 99.9 percent, while the eggs studied could be classified into the four farming methods with an accuracy of 97.1 percent. In addition, the researchers were able to determine the origin of laying hens (Lohmann Selected Leghorn, Dekalb, Lohmann Brown, Lohmann Sandy breeds) with a model accuracy of 98.4 percent.

According to the researchers, the results offer great potential for monitoring food products, for example in case of suspected cases or during random checks in retail and processing plants. Companies can order the appropriate tests and thereby reliably check whether the stamp code on the eggshell correctly indicates the type of laying hen kept.

For a method to become established, models must be supplemented with additional authentic samples and remain relevant. Models only make sense if additional, partly unknown influencing factors, such as other breeds and feeds, are taken into account. DIL therefore invites companies, research institutes and associations to work together to further develop this innovative analytics and its possible applications.