slingshot-trajectory-embedding

slingshot-trajectory-embedding

Summary

Use Slingshot trajectory analysis within ArchR to compute pseudotime and developmental lineage assignments for single-cell ATAC-seq data, inferring continuous developmental progressions from discrete cell clusters.

When to use

You have an ArchR project with clustered single-cell ATAC-seq cells and want to reconstruct developmental or cellular transition trajectories. Use this skill when your research question requires ordering cells along a developmental continuum (e.g., to infer pseudotime) and you prefer Slingshot over Monocle3 for its minimal assumptions about trajectory topology.

When NOT to use

• Your cells are not yet clustered or you lack a computed dimensionality reduction embedding — run addIterativeLSI and clustering steps first.
• You require trajectory analysis on RNA expression rather than ATAC chromatin accessibility — consider RNA-based trajectory tools or use paired scRNA-seq data with addGeneExpressionMatrix first.
• You need explicit comparison of multiple trajectory algorithms in a single analysis — use this skill for Slingshot only; pair it separately with addMonocleTrajectory if you want parallel Monocle3 results.

Inputs

• ArchR project object with computed cluster assignments
• ArchR project object with dimensionality reduction embedding (e.g., iterativeLSI or combined dims)

Outputs

• ArchR project object with Slingshot trajectory embedding and pseudotime assignments
• Trajectory lineage and pseudotime metadata per cell

How to apply

Load a prepared ArchR project object containing computed cluster assignments and dimensionality reduction embeddings (e.g., from addIterativeLSI or addCombinedDims). Invoke addSlingShotTrajectories on the ArchR project, specifying the clustering and dimensionality reduction parameters. Slingshot will fit smooth spline curves through the cluster centers, inferring pseudotime and lineage assignments for each cell. The function returns the modified ArchR project with trajectory embeddings stored internally; inspect the trajectory results to validate that inferred lineages match your expected developmental stages or biological branching pattern.