Problems with Harmony k-means calculation

DendrouLab / panpipes

Multi-modal single cell analysis pipelines

BSD 3-Clause "New" or "Revised" License

54 stars 9 forks source link

Problems with Harmony k-means calculation #239

Open wlason opened 6 months ago

wlason commented 6 months ago

It's a problem with harmonypy so I opened an issue in their repo. https://github.com/slowkow/harmonypy/issues/10#issuecomment-2010299183

I wonder if adding nclust parameter in pipeline.yaml in the integration workflow might help?

bio-la commented 6 months ago

@wlason can you put in a Pull request with this feature added? the integration CI will pick it up and test if harmony runs with the additional parameter.

olivermccallion commented 3 weeks ago

Quick comment to say I ran into this same issue. For me, I had an empty category in my protein modality. Filtering the unused category fixed the issue for me (varying nclust and using Kmeans2 v kmeans didn't)