Ignore uncategorized cells (NaN), when calculating fractions? about scirpy

[grst]

In GitLab by @grst on Feb 11, 2020, 11:18

Ideally, we make it configurable. E.g. dropna=True as default option.

In most cases it probably still makes sense to exclude them, because they are most likely other cell-types.

[grst]

In GitLab by @szabogtamas on Feb 11, 2020, 11:43

The complication here is that this might affect multiple columns and the user might decide to drop nan in one of the columns but keep them in another. For now, I would go with dropping nan by default, by letting the user supply a whitelist of cells.

[grst]

In GitLab by @grst on Feb 11, 2020, 11:49

Wouldn't dropna make mostly sense for the target_col?

If someone wants to filter specificly, he could use dropna=False and just subset the anndata accordingly before plotting:

st.pl.group_abundance(adata[cell_whitelist, :], dropna=False)

[grst]

In GitLab by @szabogtamas on Feb 11, 2020, 12:21

In the end it will sure come to this: if the user wants to visualize something for a specific subset of cell, he will have to filter himself. I think this should be ok.

[grst]

In GitLab by @grst on Mar 16, 2020, 11:51

I close this for now, to keep things simple.

[grst]

In GitLab by @grst on Mar 16, 2020, 11:51

closed