Error with large mouse dataset

[fpauler]

Error with large mouse dataset

[fpauler]

Hi,

Following up on m own issue I now tried to run the pipeline with reduced number of genes or cells resulting in the same error. I attached both of these input files here. Both are 10X data and the command I run is the following: python src/scFEA.py --data_dir data --input_dir data --test_file [one of the attached] --moduleGene_file module_gene_mouse_m162.csv --stoichiometry_matrix cmMat_mouse_c66_m162.csv --sc_imputation True
test_200_cells.csv.gz
test_500_genes.csv.gz

[fpauler]

Hi again,
So it seems the error I reported above can be solved by removing modules from module_gene_mouse_m162.csv that have no matching gene in the count matrix.
I ran the large data set (2566 cells and 17278 genes) again wit the filtered module file and now I get another error (see below).
Thanks for your help!

Starting process data...
Process data done.
Starting train neural network...
Traceback (most recent call last):
File "scFEA/src/scFEA.py", line 327, in Starting process data...
Process data done.
Starting train neural network...
Traceback (most recent call last):
File "scFEA_analysis/scFEA/src/scFEA.py", line 327, in
main(args)
File "scFEA/src/scFEA.py", line 205, in main
out_m_batch, out_c_batch = net(X_batch, n_modules, n_genes, n_comps, cmMat)
File ".conda/envs/scFEA_env/lib/python3.9/site-packages/torch/nn/modules/module.py", line 889, in _call_impl
result = self.forward(*input, **kwargs)
File "scFEA/src/ClassFlux.py", line 47, in forward
c = self.updateC(m, n_comps, cmMat)
File "scFEA/src/ClassFlux.py", line 31, in updateC
tmp = m * cmMat[i,:]
RuntimeError: The size of tensor a (158) must match the size of tensor b (162) at non-singleton dimension 1