It is my understanding that the cells whose gene expression we have to predict in crunch 2 dont have an associated xenium data in the training set. Does this mean it is mandatory to participate in crunch 1 and have a model for predicting xenium from histology and then use this model to predict xenium and then predict gene expression from the learnt mapping in crunch 2. Please let me know if i am mistaken
It is not necessary to participate in Crunch 1, but it will help a lot as you will be able to reuse your model for Crunch 2.