SEACells on Custom Dataset — Memory Issue During model.fit()

[dacon06]

Hi,

I'm currently trying to reproduce the results from the SEACell_COVID_integration.ipynb notebook using my own dataset. However, I'm encountering an issue when attempting to compute .obs['SEACell']. Specifically, running model.fit(min_iter=10, max_iter=50) leads to an out-of-memory error, despite having access to 120 GB of RAM. My dataset consists of 132706 cells × 33694 genes.

I've experimented with adjusting the number of n_SEACells used for initialization, but the problem persists. Could you share any insights or recommendations on how SEACell was applied to the COVID-19 dataset, such as preprocessing steps, parameter choices, or strategies to reduce memory usage?

Best regards