search

aertslab / scenicplus

SCENIC+ is a python package to build gene regulatory networks (GRNs) using combined or separate single-cell gene expression (scRNA-seq) and single-cell chromatin accessibility (scATAC-seq) data.
Other
186 stars 29 forks source link

Region set, scores.feather and NameError: name 'run_cgs_models_mallet #451

Open strawberry098 opened 3 months ago

strawberry098 commented 3 months ago

Do I need the region set and scores.feather files to run snakemake on scenicplus?

According to: https://pycistopic.readthedocs.io/en/latest/faqs.html It seems like 10x multiome (ATAC) Seurat object analyzed by Signac only requires the count matrix as input.

How do I generate the region set file? And the scores.feather file does not seem to download.

I tried to follow this tutorial to generate a region set file: https://github.com/aertslab/pycisTopic/blob/old/notebooks/Toy_melanoma-RTD.ipynb

But, I'm getting an error: NameError: name 'run_cgs_models_mallet' is not defined even though I installed mallet and used the following code: models=run_cgs_models_mallet(path_to_mallet_binary, cistopic_obj, n_topics=[2,5,10,15,20,25,30,35,40,45,50], n_cpu=24, n_iter=500, random_state=555, alpha=50, alpha_by_topic=True, eta=0.1, eta_by_topic=False, tmp_path=outDir + 'SCRATCH/', #Use SCRATCH if many models or big data set save_path=outDir)

SeppeDeWinter commented 3 months ago

Hi @strawberry098

Yes you will need both, please see: https://scenicplus.readthedocs.io/en/latest/tutorials.html for more information. Regarding the NameError: name 'run_cgs_models_mallet' is not defined: Did you install and import pycisTopic?

All the best,

Seppe

strawberry098 commented 3 months ago

Hi @SeppeDeWinter

Thanks for the reassurance. I'm new to python, and I realized I didn't import (despite installed) mallet. After importing, the mallet model function ran.

The reason why I opened issue #452 was because after I created the cisTopic object using a 10x Multiome snATAC-seq peak counts matrix, I didn't get the cell annotation info. I realized the peak counts matrix normally does not include cell annotation info, but how can I add it to the cisTopic object?

Thank you!

SeppeDeWinter commented 3 months ago

Hi @strawberry098

Please see: https://pycistopic.readthedocs.io/en/latest/notebooks/human_cerebellum.html#Adding-metadata-to-a-cisTopic-object for how to add cell metadata to your cisTopic object.

Seppe