b'ASSOCIATION NEWS Figure 1. Digital PCR results for pure annual rye grass (a) and pure perennial rye grass (b) varieties for one of the markers tested. FAM is the dye used for the annual rye grass SNP allele; SUN (HEX) is the dye used for the perennial rye grass SNP allele. [Data generated by Shaun Bushman and his team at USDA.]Figure 2. Digital PCR results for different contamination levels of annual rye grass in perennial rye grass varieties (2%, 4%, 6%, 10%, 20% and 50%). The intensity of the fluorescence changes from SUN (HEX) to FAM along with the different levels of ARG contamination. The example is for one of the markers tested. FAM is the dye used for the annual rye grass SNP allele; SUN (HEX) is the dye used for the perennial rye grass SNP allele. [Data generated by Shaun Bushman and his team at USDA]for the validation of markers on a larger set ofcomes as an Excel file, indicating the percentageRules, including running comparative tests varieties. of identity of the sample with a variety or theand the various DNA-based testing methods whole set of varieties learned by the software.for verification of species and varieties. The Artificial Intelligence-Based Markers In 2023, the data obtained after the first test wascontent of the handbook goes from the history The Variety Committee is running an ISTAanalysed, after considering some validationof DNA-based varietal identification within Special Project, led by Ana Laura Vicario,parameters such as sensitivity and repeatability.ISTA; guidelines for comparative test (CT) together with the Advanced TechnologiesA second data set was generated during 2024 organisation; validation of DNA-based markers Committee to use neuronal network-based(Fig. 3), and the data is currently being analysedand statistical approaches for data analysis; markers for variety identification. The softwareby the Statistics Committee to evaluate if the twoto laboratory accreditation for DNA-based from ZoomAgri can be trained to discriminatedatasets gave equivalent results. Next steps withtesting. Throughout 2024 and 2025, the working varieties that are present in a reference database.regard to the validation of this technology will begroup had many technical discussions on The equipment consists of a scanner (for takingdiscussed with the Statistics Committee. the different topics selected. The group also the seed image) and a computer (for imageconsulted with the Statistics Committee and the analysis and results output). The software learns,Handbook on DNA-Based Testing ISTA Accreditation and Technical Department. through convolutional neural networks, theLed by Marie-Claude Gagnon, the VarietyOne of the topics discussed with the Statistics morphological profile of thousands of seeds fromCommittee set up a working group for publishingCommittee was the application of a new the same variety to gather as much variability asa handbook on DNA-based testing in 2023.methodology for analysis and interpretation possible (i.e. harvesting years and climate andThe aim of the handbook is to have a guideof data from CTs. The proposed methodology soil conditions). The output from the softwarefor adding new DNA-based tests into the ISTAis better suited to the type of data produced 48 SEED TESTING INTERNATIONAL www.seedtest.org'