search

slowkow / harmonypy

🎼 Integrate multiple high-dimensional datasets with fuzzy k-means and locally linear adjustments.
https://portals.broadinstitute.org/harmony/
GNU General Public License v3.0
198 stars 22 forks source link

Error in running large dataset #32

Closed zhou-1314 closed 10 months ago

zhou-1314 commented 10 months ago

First of all, thank you very much for migrating the Harmony algorithm from R to Python. Today, I encountered the following error while integrating 318,720 cells from different datasets using harmonypy. error_information.png Hope to receive your help. Thank you.

zhou-1314 commented 10 months ago

It's solved. when I used harmony_integrate function of scanpy, it's worked.

sce.pp.harmony_integrate(adata, 'aligned_dataset',basis='X_uce')
slowkow commented 10 months ago

Thanks for sharing your experience.

If we search the web for "precompiled NUM_THREADS exceeded" we can find many results with suggestions for how to solve this issue.

For example, this Stackoverflow post suggests that we can set the NUM_THREADS option like this:

os.environ['OPENBLAS_NUM_THREADS'] = '1'

and then the sklearn.KMeans function should run without error.

I do not know whether or not this will fix your issue, because I don't know your system environment (python version, package versions, etc.).