search

yunguan-wang / Spacia

BSD 3-Clause "New" or "Revised" License
22 stars 2 forks source link

Help with running spacia, ERROR:::Spaca results don't match input! #54

Open Chrisdoan9 opened 1 week ago

Chrisdoan9 commented 1 week ago

Hi team,

I try spacia on a data with around 300,000 cells. However, I got an error, so I try on a subset with 300 cells. Seem the software ran well. However, I got this error on subset of 3,000 cells.

python spacia.py cell_by_gene_aligned.txt spacia_metadata_aligned.txt --receiver_cluster Astrocyte --sender_cluster Microglia-PVM --receiver_features CD74 --sender_features NOS1

cat spacia_log.txt 15:00:11,INFO:::Processing expression counts. 15:00:12,INFO:::Maximal distance for 10 expected neighbors is 671.38 15:00:12,INFO:::Limiting bags to those with at least 2 sender cells 15:00:12,INFO:::Number of bags: 338 15:00:12,INFO:::Cosntruct Receiver pathways from each input gene 15:00:12,INFO:::Constructing pathway using correlation aggregation 15:00:12,INFO:::Cosntruct Sender pathways from each input gene 15:00:12,INFO:::Constructing pathway using correlation aggregation 15:00:12,INFO:::Writing spacia_job.R inputs to the model_input folder. 15:00:12,INFO:::CD74_correlated_genes is already finished and will be skipped. 15:00:12,INFO:::Running spacia_R MCMC MIL models. 15:00:12,INFO:::Collecting results. 15:00:12,INFO:::Spacia_R_results at: 15:00:12,INFO::: spacia/CD74_correlated_genes 15:00:12,ERROR:::Traceback (most recent call last): 15:00:12,ERROR::: File "spacia.py", line 1051, in 15:00:12,ERROR:::assert ( 15:00:12,ERROR:::AssertionError 15:00:12,ERROR:::: 15:00:12,ERROR:::Spaca results don't match input!

Would you please have a suggestion? I checked the cell ID of cell_by_gene_aligned.txt and spacia_metadata_aligned.txt and they are the same. Thank you so much! cell_by_gene_aligned.txt spacia_metadata_aligned.txt

wtwt5237 commented 1 week ago

Hello @Chrisdoan9

Please contact Yunguan Wang, [yunguan.wang@cchmc.org] directly for your question. I believe he wrote this part of the codes

Thanks!

Tao

Chrisdoan9 commented 1 week ago

Hi @wtwt5237,

Thank you for your suggestion! I reached out on Sunday and waiting for the reply.

yunguan-wang commented 1 week ago

Generally you will need more than 500 bags to run Spacia. MCMC may not be able to converge and return meaningful data. As for the error. It is likely MCMC did not output the primary instances correctly in a expected manner. Spacia assume a one-to-one mapping between each receiver cell and each of its candidate sender cells. Can you check the content in the "pip_recal.txt" file?

Chrisdoan9 commented 1 week ago

Hi @yunguan-wang,

Thank you for your reply! I think the file pip_recal.txt is only available for a successful run. It is in the folder spacia/CD74_correlated_genes/ CD74_correlated_genes_pip_recal.txt Do I need to change any parameter or increase number of cells to get number of bag above 500?