Trajectory Lineage Agent Skill

Trajectory And Lineage

Version Compatibility

Reference examples assume recent stable releases of the preferred tools, especially scanpy and the other tools listed below.

Before using code or command patterns, verify installed versions match the environment:

• Python: python -c "import <module>; print(<module>.__version__)"
• CLI: <tool> --version
• If signatures differ, inspect the installed help or API and adapt the pattern instead of retrying unchanged.

Overview

Workflow for pseudotime, lineage branching, and state-transition analysis in single-cell data.

When To Use This Skill

• use when the user asks for pseudotime, lineage branching, or developmental progression
• use when the single-cell object already has a coherent embedding and annotations
• use when dynamic gene programs or branch-specific markers are needed

Quick Route

• If the input is raw or minimally processed data, start with validation and QC before any modeling.
• If the input is already processed, skip directly to the first workflow step that matches the user goal.
• If the user asks for a biological conclusion, always produce at least one QC or confidence artifact alongside the final result.

Progressive Disclosure

• Read references/technical_reference.md when you need deeper tool-selection rules, environment adaptation notes, or extra validation guidance.
• Keep SKILL.md as the main execution path and load the reference file only when the task or failure mode needs the extra detail.

Default Rules

• Prefer Python-first workflows unless the task explicitly requires something else.
• Keep intermediate and final outputs separated.
• Record software versions, reference builds, and key parameters when they affect interpretation.
• Favor reproducible tables and figures over one-off interactive-only outputs.

Expected Inputs

• processed single-cell object
• cluster labels
• optional time or perturbation metadata

Expected Outputs

• pseudotime assignments
• branch or lineage states
• dynamic gene programs

Preferred Tools

• scanpy
• scvelo where velocity is available
• matplotlib

Starter Pattern

Preferred starting point: scanpy
Inputs: processed single-cell object, cluster labels, optional time or perturbation metadata
Outputs: pseudotime assignments, branch or lineage states, dynamic gene programs

Workflow

1. Check topology assumptions

Ensure the embedding and cluster relationships support a trajectory-style interpretation.

2. Pick roots and branches carefully

Use prior biology or metadata to justify start states and branch structure.

3. Infer trajectories

Compute pseudotime or lineage paths and verify they align with marker trends.

4. Identify dynamic features

Report genes or modules that vary along pseudotime or across branches.

5. Visualize with context

Overlay trajectories on embeddings and summarize branch-specific biology.

Output Artifacts

• Recommended output layout:
• results/ for final tables and serialized objects
• figures/ for plots and static visual exports
• qc/ for checks that justify downstream interpretation
• Minimum expected outputs for this skill:
• pseudotime assignments
• branch or lineage states
• dynamic gene programs

Quality Review

• Confirm identifiers and metadata join correctly before modeling or summarizing.
• Generate at least one QC artifact before final biological interpretation.
• Keep raw or minimally processed inputs separate from transformed outputs.
• Review embeddings together with QC metrics and batch structure before labeling biology.
• Preserve the processed object with metadata and embeddings for downstream reuse.

Anti-Patterns

• forcing linear trajectories on clearly disconnected states
• setting roots arbitrarily without stating the assumption
• claiming lineage causality from static data alone

Related Skills

• scRNA Preprocessing And Clustering
• Cell Annotation
• Cell Communication
• Multiome And scATAC

Optional Supplements

• scvelo