AttributeError: 'csc_matrix' object has no attribute 'A'

[fusca9]

Dear Author, i'm using this script for analysis: import SEVtras SEVtras.sEV_recognizer( input_path='/mnt/data1/foschi',
sample_file='/mnt/data1/foschi/sample_file.txt',
out_path='/mnt/data1/foschi/outputs_SEVtras', species='Homo' ) the output directories of cellranger are in: /mnt/data1/foschi and the file .txt is in this path: '/mnt/data1/foschi/sample_file.txt. However, i encountered this error: The above exception was the direct cause of the following exception:

Traceback (most recent call last): File "/mnt/data1/foschi/script_SEVtras.py", line 4, in SEVtras.sEV_recognizer( File "/home/foschi/.local/lib/python3.10/site-packages/SEVtras/main.py", line 121, in sEV_recognizer inter_adata, iteration_list = iteration(inter_adata, iteration_list, thershold, max_M, threads=threads, number_g = 30, alpha = alpha) File "/home/foschi/.local/lib/python3.10/site-packages/SEVtras/utils.py", line 204, in iteration result_tmp = multi_cor(inter_adata, threads) File "/home/foschi/.local/lib/python3.10/site-packages/SEVtras/utils.py", line 173, in multi_cor result_tmp = pool.map(corr_genes, item_list) File "/usr/lib/python3.10/multiprocessing/pool.py", line 367, in map return self._map_async(func, iterable, mapstar, chunksize).get() File "/usr/lib/python3.10/multiprocessing/pool.py", line 774, in get raise self._value AttributeError: 'csc_matrix' object has no attribute 'A'. i don't know why this error occurs, i followed all the instructions on the website. Do you know what is going wrong? Can ypu please give me some advice? Thank you very much for your work!

[RuiqiaoHe]

The error may be due to mismatched format in the input data files (different from 10X CellRanger standard output). It is OK to format each dataset in the sample_file.txt with the following code: import anndata as ad from scipy.sparse import csr_matrix adata = ad.read_h5ad('./adata.h5ad') adata.X = csr_matrix(adata.X) adata.write_h5ad('./adata_s.h5ad') The generated adata_s.h5ad would be used as the input files.

[fusca9]

Thank you for the quick answer! however i used 10XCellRanger standard outputs(raw_feature_bc_matrix) as inputs for SEVtras. Anyway i'll try your suggestion!

[RuiqiaoHe]

If you used the input data from CellRanger outputs, could you please check the version of scipy in your environment? An earlier version of scipy than 1.14.0 should work:

Several previously deprecated methods for sparse arrays were removed: asfptype, getrow, getcol, get_shape, getmaxprint, set_shape, getnnz, and getformat. Additionally, the .A and .H attributes were removed.

From https://docs.scipy.org/doc/scipy/release/1.14.0-notes.html

[fusca9]

Thank you very much for the informations. I tried with an earlier version of scipy and it goes, although it takes 2 hours to analyze 3/16 samples and it's using all the threads available on the server where i work. is it normal?

[RuiqiaoHe]

This is not normal. Could you please first try to run the test data located at https://github.com/bioinfo-biols/SEVtras/tree/main/tests? It would only take a few minutes.

[fusca9]

i'm sorry for the late answer. i had a problem with th server but now it works and i could achieve finishing the SEVtras.sEV_recognizer analysis. Thank you very much. i would ask only one last question. The input for SEVtras.ESAI_calculator consists of a cell matrix with cell type, but how can i create this matrix? what kind of data i need? sorry if i'm not very into english and thank you for the help!

[RuiqiaoHe]

SEVtras needs the cell matrix in h5ad format, this can be converted by scanpy with the following code: import scanpy as sc adata = = sc.read_10x_h5(path)#path means the directory of feature_bc_matrix adata.write_h5ad('./adata.h5ad') And the cell type information, this is conventional analysis in the single cell transcriptomics. You can refer to here for further insights.