Question regarding the datasets

[hozumiyu]

I downloaded the data using the dance.datasets.singlemodality.ScDeepSortDataset
In the train/test split, there are multiple data files for the same tissues.
For instance, for mouse liver, both training and testing data have multiple mouse liver tissues.
Do these train/test data correspond to one another, or do they originate from different studies?
Is the standard practice to merge the same tissues and test out a model?

Additionally, who do I cite for these data? I may have missed it in the documentation, but I cannot find the authors who have performed the experiments.
Lastly, is there a documentation for each of the files obtained from download_all, like number of gene/cells, cell type distribution, etc?