consistency: Create, order, and plot a consistency matrix

In kieranrcampbell/ouija: Interpretable marker-based single-cell pseudotime using Bayesian parametric models

Description

The consistency matrix is the N-by-N (for N cells) matrix where the entry in the i-th row and j-th column describes the empirical probability that cell i is ordered after cell j. These functions allow the calculation, ordering (by posterior pseudotime) and plotting of the consistency matrix. See cluster_consistency to cluster the cells using Gaussian Mixture Modelling.

Usage

plot_consistency(oui, cmo = NULL, interpolate = FALSE)

consistency_matrix(oui)

consistency_matrix_ordered(oui, cmat = NULL)

Arguments

oui	A ouija_fit object
cmo	An optional ordered consistency matrix
interpolate	Passed to geom_raster
cmat	A pre-computed consistency matrix returned by consistency_matrix(oui)

Details

The consistency matrix is useful for clustering the pseudotime trajectory into discrete stages, as if there are regions where the empirical ordering probability is around 0.5 then there is large uncertainty as to the ordering, implying the cells are of roughly the same type. If the probability is closer to 0 or 1 then there is less uncertainty and the cells are probably undergoing a smooth transition.

Value

A cell-by-cell consistency matrix (ordered by posterior pseudotime if consistency_matrix_ordered), or a ggplot2 plot object displaying the ordered consistency matrix (if plot_consistency_matrix)

See Also

cluster_consistency

Examples

data(oui)
plot_consistency(oui)
cmat <- consistency_matrix(oui)

cmo <- consistency_matrix_ordered(oui, cmat)

kieranrcampbell/ouija documentation built on Dec. 2, 2019, 12:44 p.m.