ValueError: After removing rows with only NaN interactions, none remain.

[realzehuali]

Following the same jupyter notebook (tutorial_visium_hne.ipynb) with the same tutorial data but got the following error:

ValueError Traceback (most recent call last)
Cell In[10], line 6
1 sq.gr.ligrec(
2 adata,
3 n_perms=100,
4 cluster_key="cluster",
5 )
----> 6 sq.pl.ligrec(
7 adata,
8 cluster_key="cluster",
9 source_groups="Hippocampus",
10 target_groups=["Pyramidal_layer", "Pyramidal_layer_dentate_gyrus"],
11 means_range=(3, np.inf),
12 alpha=1e-4,
13 swap_axes=True,
14 )

File \squidpy\lib\site-packages\squidpy\pl_ligrec.py:295, in ligrec(adata, cluster_key, source_groups, target_groups, means_range, pvalue_threshold, remove_empty_interactions, remove_nonsig_interactions, dendrogram, alpha, swap_axes, title, figsize, dpi, save, **kwargs)
292 pvals = pvals[pvals <= pvalue_threshold]
294 if remove_empty_interactions:
--> 295 pvals, means = filter_values(pvals, means, mask=(pd.isnull(means) | pd.isnull(pvals)), kind="NaN")
296 if remove_nonsig_interactions and alpha is not None:
297 pvals, means = filter_values(pvals, means, mask=pvals <= alpha, kind="non-significant")

File ~\squidpy\lib\site-packages\squidpy\pl_ligrec.py:205, in ligrec..filter_values(pvals, means, mask, kind)
202 means = means.loc[mask_rows]
204 if pvals.empty:
--> 205 raise ValueError(f"After removing rows with only {kind} interactions, none remain.")
207 mask_cols = mask.any(axis=0)
208 pvals = pvals.loc[:, mask_cols]

ValueError: After removing rows with only NaN interactions, none remain.

Thank you for any help!

[Hagibaer]

+1

[caiquanyou]

+1

[Silaschuwen]

+1 yes the same error :(

[giovp]

hi, which version of squidpy are you using?

[Silaschuwen]

hi, which version of squidpy are you using?

1.4.1

[amonell]

I got the same error using squidpy on my own data. It happened when I was using raw counts. After log-normalizing, the function worked.

sc.pp.normalize_total(adata, target_sum=1e4)
sc.pp.log1p(adata)

[giovp]

@Silaschuwen have you used normalized data?

[Silaschuwen]

@Silaschuwen have you used normalized data?

i tried normalized data but it doesn't look like it's working out very well. here is my code (the same as the notebooks/examples)

@amonell may i ask you what is your version of squidpy?

import squidpy as sq

adata = sq.datasets.seqfish()

sc.pp.normalize_total(adata, target_sum=1e4)

sc.pp.log1p(adata)

res = sq.gr.ligrec(
    adata,
    n_perms=1000,
    cluster_key="celltype_mapped_refined",
    copy=True,
    use_raw=True,
    transmitter_params={"categories": "ligand"},
    receiver_params={"categories": "receptor"},
)

sq.pl.ligrec(res, source_groups="Erythroid", alpha=0.005)

ValueError: After removing rows with only NaN interactions, none remain.

[amonell]

@Silaschuwen You have "use_raw" set to true which may be using un-normalized counts. I would try setting it to False.
Maybe the default threshold is too high, you can try setting "threshold = 0".

I am using Squidpy version 1.3.1 in python 3.10

[Silaschuwen]

@Silaschuwen You have "use_raw" set to true which may be using un-normalized counts. I would try setting it to False. Maybe the default threshold is too high, you can try setting "threshold = 0".

I am using Squidpy version 1.3.1 in python 3.10

thanks so much! everything works smoothly~