snapatac2.pl.spectral_eigenvalues#

snapatac2.pl.spectral_eigenvalues(adata, width=600, height=400, show=True, interactive=True, out_file=None)[source]#

Plot spectral embedding eigenvalues and mark the elbow.

Use this function after spectral decomposition to choose the number of eigenvectors retained for downstream analysis.

Anti-Patterns#

Do NOT call this before computing spectral eigenvalues. The input must contain adata.uns["spectral_eigenvalue"].
Do NOT treat this as a pure plotting helper; it also writes the inferred elbow to adata.uns["num_eigen"].

param adata:: Annotated data matrix containing uns["spectral_eigenvalue"].
type adata:: AnnData
param width:: Width of the rendered plot in pixels.
type width:: int
param height:: Height of the rendered plot in pixels.
type height:: int
param show:: Whether to display the figure immediately.
type show:: bool
param interactive:: Whether to display an interactive Plotly figure when show=True.
type interactive:: bool
param out_file:: Output path for saving the plot. Supported suffixes include .svg, .pdf, .png, and .html.
type out_file:: str | None
returns:: Returns a Plotly figure when show=False and out_file=None; otherwise renders or saves the plot and returns None.
rtype:: plotly.graph_objects.Figure’ | None

Examples

>>> import numpy as np
>>> import snapatac2 as snap
>>> adata = snap.AnnData(X=np.ones((3, 3)))
>>> adata.uns["spectral_eigenvalue"] = np.array([4.0, 2.5, 1.2, 0.4])
>>> fig = snap.pl.spectral_eigenvalues(adata, show=False)
>>> fig.update_layout(title="Spectral eigenvalues")