ZeroDivisionError: float division by zero

[Simran-Sodhi]

Hi,

I have been getting this error: ZeroDivisionError: float division by zero
when trying to run the following code:
de.test.wald( data=adata_human, gene_names=adata_human.var_names, formula_loc="~ 1 + palantir_pseudotime_int", factor_loc_totest="palantir_pseudotime_int", as_numeric=["palantir_pseudotime_int"] )
Here is the complete error log: zero_division_error.pdf
I am getting the same error even when I use a discrete variable instead of a continuous one and when I use sample single-cell data from the Scanpy datasets (scanpy.datasets.ebi_expression_atlas(accession = "E-GEOD-98816"))
`
Current Versions:

• numpy==1.21.6
• anndata==0.8.0
• pandas==1.4.3
• scanpy==1.9.1
• diffxpy.api==v0.7.4+21.g12f1286

Could you please look into this? Thanks for your help

[mhgh146]

I've run into the same error when running de.test.versus_rest. Iterations near end and then generates the same above error.

[guangjunyin]

The same error, hoping an answer.

[victorsanchezarevalo]

The same error, any clue?

[guangjunyin]

The same error, any clue?

[guangjunyin]

The same error, any clue?
Here is the detailed information.

[victorsanchezarevalo]

I had exactly the same.

[theGreen-Coder]

Same here.

[theGreen-Coder]

It's weird because the same code works on Collab but not on my computer. Collab: https://colab.research.google.com/drive/1FUZXHNCXtqWlip0jQN1Z9QBJ7JBO5qIR?usp=sharing
When I try to run a wald test:

test = de.test.wald(
    data=data,
    formula_loc="~ 1 + condition",
    factor_loc_totest="condition",
)

# Error Terminates the program here on my Mac

summaryTest = test.summary().iloc[:10,:]

It works on Google Collab, but not on my MacOS. When I run it on Google Collab, it also gives me an error (below), but that doesn't stop the execution of my code. And therefore, test is defined, as well as summaryTest.
However, when I run this exact same code on my Mac machine, this happens (below). An exception is thrown and the whole program stops (leaving test undefined, thus leaving also summaryTest undefined).

Error on Google Collab (last line)

Error on my Mac

The only way this works on my Mac is if I ran this line of code before the wald test:

sc.pp.highly_variable_genes(data, n_top_genes=255, subset = True, flavor = "seurat_v3", batch_key="sample")

[Alixbonard]

Hi,

I have been facing the same problem. Did someone find a solution ?

[pdrofelipe]

Hello,

same problem here as well.
It would be nice an answer trying to shed a light here

[Alixbonard]

Hello,
I found the solution.
The problem is related to the dask package. I downgraded it to dask==2021.4.0.
Normally with this version the error becomes a warning.
N.B. I use Diffxpy==0.7.4 and batchglm==0.7.4.

[DieTANG]

Hi, I have the same error, any solution so far?

anndata 0.9.1
diffxpy v0.7.4
matplotlib 3.7.1
numpy 1.24.3
pandas 2.0.2
scanpy 1.9.3
scvi 1.0.0
seaborn 0.12.2
session_info 1.0.0
IPython 8.14.0
jupyter_client 8.2.0
jupyter_core 5.3.0
notebook 6.5.4
Python 3.9.16 | packaged by conda-forge | (main, Feb 1 2023, 21:39:03) [GCC 11.3.0]
Linux-6.2.12-arch1-1-x86_64-with-glibc2.37

[Alixbonard]

Hi @DieTANG,

Did you try to downgrade dask ? dask==2021.4.0 worked for me.

[Marwansha]

i have the same issue here, near convergence it give this error, anyone managed?
i already have dask==2021.4.0