Hello,
We have realized that there are problems with the genotyping data that covers trial AWY1_DVPWA_2024. We have added new genotype data for this trial that should be easier to work with. That data can be found by searching for genotype protocol: GBS WSU 2023 and genotype project: WSU_2023_GBS (and the link to the data on the predictathon rules page has been updated).
Clearly, there have always been two important challenges to being successful in the Predictathon: A) Skill with the T3 database to either find and download data from the GUI or through BrAPI, and B) Skill at using data to make predictions. We became concerned that Skill A would be too great a barrier for many participants. Our belief is that the best use of T3 can come from a combination of Skills A and B. Nevertheless, for this first Predictathon, we wanted to make it possible for scientists who already have Skill B but not the time to develop Skill A to participate. To this end, we have identified T3 data that could be reasonably applied to predicting the Predictathon trials using simple heuristics:
- We identified phenotyping trials that evaluated some of the same accessions as each Predictathon trial. The assumption is that the overall germplasm in those trials will be related to the Predictathon accessions.
- We identified archived VCFs with genotype data from a minimum number of each of the accessions in the putative training trials.
Links to this data and metadata are now available on the Predictathon rules page. Note that these links go to data that is, in principle, static: it will not update if we get new, relevant data added to T3/Wheat. The scripts that were used to assemble the data are in a public GitHub repository: https://github.com/jeanlucj/T3_predictathon_find_training_trials also linked to the Predictathon rules page. We will update the data for the last time around March 20th.
Because of these issues, we are extending the deadline for the Predictathon from March 13th to March 27th.
- The Triticeae Toolbox
