msmu.pl.plot_pca

Plot previously computed principal-component scores.

Parameters:

Name	Type	Description	Default
mdata	MuData	MuData containing the selected modality and observation metadata.	required
modality	str	Name of the modality to plot, such as "protein" or "psm".	'protein'
groupby	str | None	Grouping column in observation metadata (or feature metadata where supported). None uses the resolved obs_column.	None
colorby	str | None	Observation column used to assign sample colors when groupby resolves to obs_column; ignored for other groupings. None uses the template palette.	None
template	str	Registered plotly template name, normally "msmu".	DEFAULT_TEMPLATE
pcs	tuple[int, int] | list[int]	Two 1-based component numbers, for example (1, 2).	(1, 2)
obs_column	str | None	Sample identifier in mdata.obs. If omitted, uses .uns["plotting"]["default_obs_column"], then source name/source_name/sample/filename, then the observation index.	None
key	str	Key containing PCA scores in modality .obsm and variance_ratio in .uns.	'X_pca'
kwargs	str	Additional plotly layout options, for example width=800 or title_text="QC".	{}

Returns:

Type	Description
Figure	plotly Figure. Call .show() to display or .write_html("plot.html") to export.

Notes

Run pca first. Match key to its key_added; at least the requested components must exist. Quantification and embeddings are unchanged. Resolving an existing sample-identifier column can convert that column in mdata.obs to categorical; pass a copy to preserve its dtype.

Examples:

import msmu as mm
mdata = mm.tl.pca(mdata, modality="protein", n_components=2)
fig = mm.pl.plot_pca(mdata, modality="protein", pcs=(1, 2))
fig.show()