Hi, thank you for releasing Nicheformer — it’s a very impressive and useful framework.
I have a question regarding the Ensembl gene ID versions used in the released Nicheformer models (e.g. the model.h5ad file).
Which Ensembl release / genome build are the ENSG (human) and ENSMUSG (mouse) gene IDs in the model based on?
This information would be very helpful for properly mapping our own mouse scRNA-seq data to the gene vocabulary expected by Nicheformer and to avoid unintended gene dropouts during tokenization.
Thank you very much for your help.
Hi, thank you for releasing Nicheformer — it’s a very impressive and useful framework.
I have a question regarding the Ensembl gene ID versions used in the released Nicheformer models (e.g. the model.h5ad file).
Which Ensembl release / genome build are the ENSG (human) and ENSMUSG (mouse) gene IDs in the model based on?
This information would be very helpful for properly mapping our own mouse scRNA-seq data to the gene vocabulary expected by Nicheformer and to avoid unintended gene dropouts during tokenization.
Thank you very much for your help.