omicverse.single.Monocle

class omicverse.single.Monocle(adata)

Monocle2-style single-cell trajectory analysis.

Wraps a pure-Python implementation of Monocle 2 as a stateful analyzer operating on an AnnData object. All results are stored in the AnnData (.obs, .var, .uns['monocle'], .obsm) so the usual scanpy workflow continues to work seamlessly.

Parameters:

adata (AnnData) – Annotated data matrix (cells × genes). Expression matrix in adata.X should be raw/normalized counts (negative binomial model).

adata

The annotated data matrix with analysis results stored in-place.

Type:

AnnData

Examples

Basic trajectory analysis:

>>> mono = ov.single.Monocle(adata)
>>> mono.preprocess()              # size factors + dispersions
>>> mono.select_ordering_genes()   # high-variance gene selection
>>> mono.reduce_dimension()        # DDRTree
>>> mono.order_cells()             # assign pseudotime + State
>>> mono.plot_trajectory(color_by='clusters')
>>> mono.plot_genes_in_pseudotime(['Ins1', 'Gcg'])

Differential expression along pseudotime:

>>> de = mono.differential_gene_test()
>>> beam = mono.BEAM(branch_point=1)

__init__(adata)

Initialise the Monocle analyser.

Parameters:

adata (AnnData) – Annotated data matrix (cells × genes). The expression matrix in adata.X should contain raw or normalised counts — the negative binomial model used by size-factor estimation and BEAM assumes count-like data. All downstream results (Pseudotime, State, dispersions, DDRTree reduction) are written back to the same AnnData, so subsequent scanpy-style workflows continue to work.

Notes

The external monocle2_py backend is imported lazily inside this constructor (not at module scope) to keep omicverse.single free of top-level ..external imports, per the architecture test.

Methods

BEAM([branch_point, branch_states, ...])	Branched Expression Analysis Modelling.
__init__(adata)	Initialise the Monocle analyser.
cal_ABCs([branch_point])	Compute the Area Between Curves for branch-specific genes.
cal_ILRs([branch_point, return_all])	Compute the Intrinsic Log-Ratio (per-gene lineage bias).
cluster_cells([method, k, ...])	Cluster cells on the reduced-dim space and write labels to adata.obs['Cluster'].
cluster_genes(expression_matrix, k[, method])	Cluster genes by their expression pattern along pseudotime.
detect_genes([min_expr])	Flag genes expressed above a threshold.
differential_gene_test([...])	Pseudotime-dependent differential expression via a likelihood-ratio test between a full GLM (gene ~ f(Pseudotime)) and a reduced null model.
dispersion_table()	Return the per-gene dispersion table populated by estimate_dispersions(), as a pandas.DataFrame.
estimate_dispersions([min_cells_detected, ...])	Fit per-gene dispersions under the negative-binomial model.
estimate_size_factors([method, round_exprs])	Estimate per-cell size factors.
fit_model([modelFormulaStr, relative_expr, ...])	Fit a per-gene GLM under the NB model.
gen_smooth_curves([new_data, trend_formula, ...])	Predict smoothed expression trajectories from a fitted model.
order_cells([root_state, reverse, ...])	Order cells along the learned trajectory, assigning Pseudotime and State.
plot_cell_clusters([color_by])	Plot cells colored by cluster in reduced-dim space.
plot_cell_trajectory([color_by])	plot_cell_trajectory — main DDRTree trajectory plot.
plot_complex_cell_trajectory([color_by])	Dendrogram-style trajectory layout (Pseudotime on Y-axis).
plot_genes_branched_heatmap([branch_point])	Heatmap of branch-specific gene expression.
plot_genes_branched_pseudotime(genes[, ...])	Gene expression split by branch.
plot_genes_in_pseudotime(genes, **kwargs)	Gene expression vs pseudotime with smoothed curves.
plot_genes_jitter(genes[, grouping])	Jitter plot of gene expression by group.
plot_genes_violin(genes[, grouping])	Violin plot of gene expression by group.
plot_multiple_branches_heatmap(branches, ...)	Multi-branch expression heatmap.
plot_multiple_branches_pseudotime(genes, ...)	Multi-branch gene expression curves.
plot_ordering_genes(**kwargs)	Dispersion vs mean-expression plot, highlighting ordering genes.
plot_pc_variance_explained([max_components])	Plot variance explained by principal components.
plot_pseudotime_heatmap([genes])	Heatmap of gene expression sorted by pseudotime.
plot_rho_delta(**kwargs)	Plot rho vs delta for density-peak clustering.
plot_trajectory([color_by])	plot_cell_trajectory — main DDRTree trajectory plot.
plot_trajectory_overlay(ax, **kwargs)	Overlay the DDRTree principal-graph skeleton + branch points on an externally-drawn axes (e.g. one produced by ov.pl.embedding(basis='X_DDRTree')).
plot_trajectory_with_embedding([color, ...])	Convenience combo: render ov.pl.embedding then overlay the DDRTree backbone + branch points.
preprocess([min_expr, verbose])	Run detect_genes(), estimate_size_factors() and estimate_dispersions() in sequence.
reduce_dimension([max_components, ...])	Reduce dimensionality and learn the principal graph.
relative2abs([method, ...])	Census normalisation — convert TPM / FPKM-scaled counts into estimated absolute transcript counts per cell.
select_ordering_genes([genes, ...])	Select genes used for trajectory inference.
set_ordering_filter(genes)	Explicitly set the list of ordering genes.

Attributes

Y	Tree-center coordinates (dim × K).
Z	Reduced-dim cell coordinates (dim × N).
branch_points	Branch-point vertex names in the learned tree.
pseudotime	Per-cell pseudotime (after order_cells).
state	Per-cell state (after order_cells).