[phlower] Error in ddhodge

[jpark27]

Hi, phlower team! Thank you so much for sharing such nice tool with community.

I come across with following error when running phlower with in-house multiome data (68K cells) and wonder what would cause the error? Maybe too many cells? Is there any chance I can run on whole cells?

Code:
# ddhodge phlower.ext.ddhodge(adata, basis='X_pca', roots=list(adata.obs['group']=='Naive'), k=4,npc=40, ndc=30,s=1)
Error:
AssertionError: "neato" with args ['-Tdot', '/tmp/tmpect13hs1/tmpu1ysizi5'] returned code: -11

best wishes,
J

[chengmingbo]

Hi,
Sorry for the late reply — I was busy with my defense.

It’s our honor that you are trying PHLOWER.

It looks like the graphviz library is not working properly in your environment.
Could you please provide a bit more information so we can diagnose?

• your OS / platform

• graphviz version

• the output of session_info.show() (so we can see your python package versions)

Also — have you tried running the example notebook?
https://phlower.readthedocs.io/en/latest/notebooks/fib2neuron.html

This can help confirm whether all dependencies are installed correctly.

Regarding the cell number — PHLOWER can be very slow with too many cells.
Could you try subsetting to ~10,000–20,000 cells?

Please keep in mind: the parameter k likely needs to be increased if phlower.ext.ddhodge results in a disconnected graph.

We also provide a helper function for subsetting:

phlower.tl.subset_celltype(adata, slot='grouping', minn=2000, ratio=0.2, seed=2024)

Here, minn sets the minimum number of cells — if a cell type is smaller than this, sub-sampling will skip that cell type, and ratio defines the proportion of cells to keep during subsetting.