Hi,

Thank you for your great work.

I tested the released ST decoder on TENX152. The decoder output has shape [N, 2000], but I am unsure how these 2000 dimensions correspond to specific genes. Currently, I can only interpret them as gene_0, gene_1, ..., gene_1999.

Could you please clarify:

1. Where can I find the gene list and gene ordering used by the decoder?
2. Which dataset was used to train the decoder?
3. Is the released decoder expected to work directly on TENX152, or does it require dataset-specific gene alignment/preprocessing?

I am currently obtaining a very low mean PCC and would like to verify whether I am using the correct gene mapping.
I have attached a screenshot of the evaluation results for reference.
Thank you very much.