@pchiang5 thanks for your message - it's hard for me to say without more information. Can you provide some details about the # of cells / features expected based on the TileMatrix as observed in the ArchR API? Do you have success when running the same function but for the GeneScoreMatrix? I.e.:

adata = ArchR_h5ad.read_arrow('S5.arrow', use_matrix="GeneScoreMatrix")

Hi @mvinyard

Using GeneScoreMatrix also encountered a similar error (please see below). Because the arrow file was created with custom genome/annotation files of both human (chr) and mouse (mchr) sources. I wonder if the unconventional mchr was the culprit of these errors because the chromosome to read shall be > 23 and the actual tile/genescore var_names in the AnnData files were around half of those in the arrow file.

adata = ArchR_h5ad.read_arrow('S5.arrow', use_matrix="TileMatrix")

Reading ArchR TileMatrix to AnnData
Chromosomes: 0%| | 0/23 [00:00<?, ?it/s]
Traceback (most recent call last):
File "", line 1, in
File "/home/pc/miniconda3/envs/snap/lib/python3.10/site-packages/ArchR_h5ad/_main/_read_arrow_to_adata.py", line 66, in _read_arrow_to_adata
arrow.to_adata(use_matrix=use_matrix, write_h5ad=write_h5ad)
File "/home/pc/miniconda3/envs/snap/lib/python3.10/site-packages/ArchR_h5ad/_main/_Arrow.py", line 52, in to_adata
self._adata = _compose_anndata(DataDict=self._DataDict,
File "/home/pc/miniconda3/envs/snap/lib/python3.10/site-packages/ArchR_h5ad/_compose_adata/_compose_anndata.py", line 27, in _compose_anndata
adata = _add_obs_var(adata, metadata, feature_df)
File "/home/pc/miniconda3/envs/snap/lib/python3.10/site-packages/ArchR_h5ad/_compose_adata/_add_obs_var.py", line 49, in _add_obs_var
adata.var = _add_var(feature_df)
File "/home/pc/miniconda3/envs/snap/lib/python3.10/site-packages/anndata/_core/anndata.py", line 882, in var
self._set_dim_df(value, "var")
File "/home/pc/miniconda3/envs/snap/lib/python3.10/site-packages/anndata/_core/anndata.py", line 798, in _set_dim_df
value_idx = self._prep_dim_index(value.index, attr)
File "/home/pc/miniconda3/envs/snap/lib/python3.10/site-packages/anndata/_core/anndata.py", line 810, in _prep_dim_index
raise ValueError(
ValueError: Length of passed value for var_names is 11627605, but this AnnData has shape: (1590, 6062095)

adata = ArchR_h5ad.read_arrow('S5.arrow', use_matrix="GeneScoreMatrix")

Reading ArchR GeneScoreMatrix to AnnData
Chromosomes: 0%| | 0/23 [00:00<?, ?it/s]
Traceback (most recent call last):
File "", line 1, in
File "/home/pc/miniconda3/envs/snap/lib/python3.10/site-packages/ArchR_h5ad/_main/_read_arrow_to_adata.py", line 66, in _read_arrow_to_adata
arrow.to_adata(use_matrix=use_matrix, write_h5ad=write_h5ad)
File "/home/pc/miniconda3/envs/snap/lib/python3.10/site-packages/ArchR_h5ad/_main/_Arrow.py", line 52, in to_adata
self._adata = _compose_anndata(DataDict=self._DataDict,
File "/home/pc/miniconda3/envs/snap/lib/python3.10/site-packages/ArchR_h5ad/_compose_adata/_compose_anndata.py", line 27, in _compose_anndata
adata = _add_obs_var(adata, metadata, feature_df)
File "/home/pc/miniconda3/envs/snap/lib/python3.10/site-packages/ArchR_h5ad/_compose_adata/_add_obs_var.py", line 49, in _add_obs_var
adata.var = _add_var(feature_df)
File "/home/pc/miniconda3/envs/snap/lib/python3.10/site-packages/anndata/_core/anndata.py", line 882, in var
self._set_dim_df(value, "var")
File "/home/pc/miniconda3/envs/snap/lib/python3.10/site-packages/anndata/_core/anndata.py", line 798, in _set_dim_df
value_idx = self._prep_dim_index(value.index, attr)
File "/home/pc/miniconda3/envs/snap/lib/python3.10/site-packages/anndata/_core/anndata.py", line 810, in _prep_dim_index
raise ValueError(
ValueError: Length of passed value for var_names is 115928, but this AnnData has shape: (1590, 60058)

@pchiang5 Thanks for this additional info. I think you are right; the unconventional concatenated human/mouse organization is most likely causing the issue. I see two options:

1. separate the objects in ArchR before conversion
2. supplied with an example dataset that reproduces the error, I could look into extending the tool to include such a case. Not sure if this is trivial or more involved without further investigation. Let me know what you think

@mvinyard

Thank you for your willingness to help. I sent you the dataset in the previous reply. Please let me know if you miss the link.