ℹ️ Tutorials     ℹ️ Reference documentation

Metrics to compare two trajectories

This R package implements several metrics for comparing two single-cell trajectories.

These include:

• Specific metrics, metrics which look at the similarity of a specific part of the trajectory, such as the topology or the cellular ordering

Name	Long name	Category
cordist	Geodesic distance correlation	cell positions
MSErf	Random Forest MSE	neighbourhood
NMSErf	Random Forest Normalised MSE	neighbourhood
R2rf	Random Forest R²	neighbourhood
NMSElm	Linear regression Normalised MSE	neighbourhood
MSElm	Linear regression MSE	neighbourhood
R2lm	Linear regression R²	neighbourhood
edgeflip	Edge flip	topology
HIM	Hamming-Ipsen-Mikhailov similarity	topology
Isomorphic	isomorphic	topology
F1branches	Overlap between the branches	branch assignment
F1milestones	Overlap between the milestones	branch assignment

• Application metrics, which assess the accuracy of some downstream analyses of trajectories

Name	Long name
corfeatures	Feature importance correlation
wcorfeatures	Feature importance weighted correlation
ksfeature	Feature importance enrichment ks
wilcoxfeature	Feature importance enrichment wilcox

• Overall metrics, which combine several specific and/or application metrics to analyse the overall similarity of two trajectories

Name	Long name
meanharmonic	Harmonic mean
meangeometric	Geometric mean
meanarithmetic	Arithmetic mean

Latest changes

Check out news(package = "dyneval") or NEWS.md for a full list of changes.

Recent changes in dyneval 0.9.9 (07-04-2019)

• MINOR CHANGE: Update categories of metrics

• MINOR CHANGE: Move metric functions from dyneval to dynutils.

• MINOR CHANGE: Bump version number for dynwrap 1.0.0

Recent changes in dyneval 0.2.2 (21-11-2018)

• DOCUMENTATION: Update and expand documentation.

• BUG FIX: Update dyneval for dynwrap 0.3.1 (#50).

• MINOR CHANGE: Clean up imports and suggests.

Dynverse dependencies