---

name: analysis-workflows description: "Run MDAnalysis analysis classes, custom AnalysisBase workflows, numerical distance analyses, and analysis backend validation." disable-model-invocation: true metadata: disco-role: operating license: NOASSERTION

MDAnalysis Analysis Workflows

Use this sub-skill when the task is to run built-in MDAnalysis.analysis modules, interpret analysis.results, convert frame loops into AnalysisBase, or debug analysis backends and numerical outputs.

Route First

• For loading files, building synthetic Universe objects, attaching trajectories, or writer setup, read ../universe-io/SKILL.md first.
• For atom selection language, empty selections, updating selections, topology attributes, masses, charges, bonds, or custom attributes, read ../selections-topology/SKILL.md first.
• For on-the-fly transformations, PBC unwrapping/centering before analysis, or writing transformed trajectories, read ../transformations-writing/SKILL.md first.

Use This For

• Running AnalysisBase.run(start=..., stop=..., step=...) or run(frames=...) and validating frames, times, n_frames, and results.
• Choosing and interpreting RMSD, RMSF, alignment, contacts, distance arrays, RDF, hydrogen bonds, PCA, MSD, leaflet, and density analyses.
• Writing a small custom AnalysisBase subclass with deterministic result arrays and optional split-apply-combine aggregation.
• Selecting safe analysis backends: default serial execution, supported multiprocessing or dask backends, and distance-function backends.
• Diagnosing analysis-specific failures such as empty selections, mass mismatches, missing charges/bonds, PBC boxes, deprecated result aliases, and result-shape confusion.

References

• references/analysis-api.md summarizes verified signatures, result containers, result shapes, and backend rules.
• references/recipes.md gives practical workflows for built-in analyses and custom AnalysisBase implementations.
• references/troubleshooting.md maps common analysis symptoms to concrete recovery steps.
• scripts/analysis_smoke_check.py runs a tiny synthetic custom analysis and distance-array validation without repo test data.

Safe Validation

Run the bundled smoke check in an environment where MDAnalysis is installed:

python scripts/analysis_smoke_check.py

Expected output includes PASS custom AnalysisBase frames/results, PASS PBC distance_array preallocated result, and PASS frames/start conflict validation.