30k genes were found in the adata. Note that Celloracle is intended to use around 1000-3000 genes, so the behavior with this number of genes may differ from what is expected.
Opened this issue · 0 comments
Hi! thank you for the tool.
I just wanted to ask a question about this warning.
30k genes were found in the adata. Note that Celloracle is intended to use around 1000-3000 genes, so the behavior with this number of genes may differ from what is expected.
I am aware the Celloracle should be used on the dataset that contains only the highly variable genes.
However, all the genes for which I am interested on simulating the KO are not in fact HVGs.
I was wondering if the "problem" of using a bigger dataset is just that is more computational expensive, (which I can solve using more resources) or if it is a conceptual error/problem.
If the latter is true, could you please elaborate more on why is it not good to use more than 3000 genes?
Thank you very much for explaining it
Best regards,
Maria