Creating scenicplus_obj

[Umaarasu]

When I try to create the scenicplus obj ..I get the error Exception: No cells found which are present in both assays, check input and consider using bc_transform_func!. I checked the inputs and the barcodes match. Has someone had a similar issue?

I converted my seurat object for scRNA to Anndata for scATAC I followed the steps given here https://scenicplus.readthedocs.io/en/latest/faqs.html#i-have-an-analysis-with-another-tool-e-g-signac-archr-can-i-still-use-scenic

Code:

from scenicplus.scenicplus_class import create_SCENICPLUS_object
import numpy as np
scplus_obj = create_SCENICPLUS_object(
GEX_anndata = adata,
cisTopic_obj = cistopic_obj,
menr = menr,
bc_transform_func = lambda x: x #function to convert scATAC-seq barcodes to scRNA-seq ones
)

Error:
Exception Traceback (most recent call last)
Cell In[108], line 3
1 from scenicplus.scenicplus_class import create_SCENICPLUS_object
2 import numpy as np
----> 3 scplus_obj = create_SCENICPLUS_object(
4 GEX_anndata = adata,
5 cisTopic_obj = cistopic_obj,
6 menr = menr,
7 bc_transform_func = lambda x: x #function to convert scATAC-seq barcodes to scRNA-seq ones
8 )
9 scplus_obj.X_EXP = np.array(scplus_obj.X_EXP.todense())
10 scplus_obj

File ~/scenicplus/src/scenicplus/scenicplus_class.py:533, in create_SCENICPLUS_object(GEX_anndata, cisTopic_obj, menr, multi_ome_mode, nr_metacells, nr_cells_per_metacells, meta_cell_split, key_to_group_by, imputed_acc_obj, imputed_acc_kwargs, normalize_imputed_acc, normalize_imputed_acc_kwargs, cell_metadata, region_metadata, gene_metadata, bc_transform_func, ACC_prefix, GEX_prefix)
530 common_cells = list(set(GEX_cell_names) & set(ACC_cell_names))
532 if len(common_cells) == 0:
--> 533 raise Exception(
534 "No cells found which are present in both assays, check input and consider using bc_transform_func!")
536 # impute accessbility if not given as parameter and subset for HQ cells
537 if imputed_acc_obj is None:

Exception: No cells found which are present in both assays, check input and consider using bc_transform_func!

[SidG13]

Not sure if this will help you here, but check out https://github.com/aertslab/scenicplus/issues/90. Generally you have a sample name which needs to be removed from one of the barcodes using the bc_transform_func.

[Umaarasu]

Thanks for the suggestion.

My issue is the GEX data is imported from seurat. Hence I do not have the adata.raw, and for ATAC data I imported them from signac and processed them in cisTopic package as suggested in scenic FAQs. I am worried maybe this is difference is causing this issue.

This is how my data looks and I can see the barcodes are similar and I don't get why the error shows up.

AnnData object with n_obs × n_vars = 3448 × 36601
obs: 'nCount_RNA', 'nFeature_RNA', 'percent.mt'
var: 'name'

print(adata.obs['nCount_RNA'])
AAACATGCAAAGGTAC-1 76.0
AAACCAACACAACCTA-1 25.0

print(cistopic_obj.cell_data)

                                                cisTopic_log_nr_acc    Cell_type        cisTopic_log_nr_frag  \

AAACATGCAAAGGTAC-1 1.0000004 EC 4.1889
AAACCAACACAACCTA-1 1.6000078 ADl 4.003676

[SeppeDeWinter]

Hi

Could you show the output of

len(set(cistopic_obj.cell_names) & set(adata.obs_names))

Best,

Seppe

[Umaarasu]

@SeppeDeWinter ..
I ran this
" len(set(cistopic_obj.cell_names) & set(adata.obs_names))" and got 0.

When I checked 'cistopic_obj.cell_names'..it had the barcodes AAACATGCAAAGGTAC-1__cisTopic while my 'adata.obs_names' barcodes where AAACATGCAAAGGTAC-1. I think this because I imported both my RNA and ATAC data from seurat, but processed the ATAC through cisTopic modelling (there was addition of __cisTopic to the barcodes).

I added __cisTopic to my barcodes at 'adata.obs_names'. That fixed the issue.
"adata.obs_names = adata.obs_names + "___cisTopic""

P.S: It would be helpful if there is a tutorial showing how to import RNA data from a seurat object other than loom file. And what to expect when ATAC data is processed through cisTopic modelling and merged with RNA for SCENIC+.

Thanks!

[yojetsharma]

I was getting this error, so i changed the barcodes in my RNA h5ad file to match them in the cistopic_obj and around 51178 common barcodes intersected.

cistopic_barcodes:
 'AGTAACGAGCCTGGTA-1:WT___WT',
 'AGGAACCAGCGGATAA-1:WT___WT',
 'AGCGGATAGCAGGCCT-1:WT___WT',
 'TAGTAACGTACTGATG-1:WT___WT',
 'CAGGCTGGTCACCAAA-1:WT___WT',
 'GGGTTTGTCCATCAGG-1:WT___WT',
 'TGGTGATTCGCTAGCA-1:WT___WT',
 ...]
 snRNA_barcodes:
 Index(['AAACAGCCAAATATCC-1:WT___WT', 'AAACAGCCACAATGTT-1:WT___WT',
       'AAACAGCCAGTATGTT-1:WT___WT', 'AAACAGCCATAATCGT-1:WT___WT',
       'AAACAGCCATCCAGGT-1:WT___WT', 'AAACATGCAACTAGCC-1:WT___WT',
       'AAACATGCAACTGGGA-1:WT___WT', 'AAACATGCAAGCGATG-1:WT___WT',
       'AAACATGCACGAATCC-1:WT___WT', 'AAACATGCAGGCCATT-1:WT___WT',

How can I pass the bc_transform function such that no transformation takes place?
I tried "lambda x: x" or "x" but that didn't solve the issue either.

[SeppeDeWinter]

Hi @yojetsharma

Indeed, lambda x: x should do the trick.

Can you share the full stack trace including command you are running so we can troubleshoot further?

All the best,

Seppe