By name: stereo-seq-publication-plotting description: Use when Stereo-seq or STOmics analysis needs paper-quality figures, spatial maps, dot plots, heatmaps, marker panels, module-score maps, legends, palettes, Arial typography, non-overlapping labels, or reuse of plotting code from public Stereo-seq article repositories.
Stereo-seq Publication Plotting
Use This For
- Turning Stereo-seq coordinates, annotations, gene scores, cell-type abundances, domains, or interaction scores into manuscript-ready figures.
- Reusing high-quality plotting templates from public Stereo-seq article code instead of writing figure code from scratch.
- Fixing figure issues such as tiny fonts, legends covering data, unstable palettes, non-equal spatial aspect, or raster-only output.
Default Requirements
- Use bundled article-derived scripts in
scripts/before writing new plotting code or searching external repositories. - Read source_code.md when deciding which template to adapt. Let the current dataset, plot type, tissue geometry, and task similarity guide the choice; do not hard-code tissue-to-tool rules.
- If no curated plotting entry fits, search code_candidates.tsv for additional article-linked figure scripts and reusable files before external search.
- Before running Python or R, inspect the local environment. Prefer
conda run -n stereo-skills-py python ...for Python plotting scripts andconda run -n stereo-skills-r Rscript ...for R plotting scripts. If a required package is missing, stop that plotting step and tell the user which package is missing and which figure is blocked. - Use Arial where available. For manuscript panels, keep axis text, legend text, and labels at least 9 pt by default; use 10-11 pt for legend titles, panel titles, and important labels.
- Export vector PDF plus 300 dpi raster when useful.
- Keep legends outside the data area where possible, and never let legends or colorbars cover the tissue/map content.
- In the final response, state the reused paper, DOI, code repository or code DOI, original file, and dataset-specific edits.
Workflow
- Identify the figure contract: input file/object, coordinate columns, grouping or value column, plot type, expected panel size, and output filenames.
- Read source_code.md and choose the closest template by paper-code similarity.
- Adapt only dataset-specific parts: file paths, column names, marker lists, labels, palette entries, bin/cell size, and output names.
- Check that the result has equal spatial aspect, readable text, and no legend or colorbar overlap.
- Report provenance and modifications.
Reusable Article Code
scripts/stereo_spatial_panel_template.py: categorical/continuous spatial-map template adapted from P09 layer maps, Endo.R spatial scatter patterns, and GF/SPF cecum marker spatial maps.scripts/stereo_dotplot_template.R: marker/program dotplot template adapted from Endo.R marker dotplots and GF/SPF cecum function dotplots.scripts/stereo_marker_spatial_grid_template.py: marker spatial grid template adapted from P09, Endo.R, and avian optic tectum marker-map figures.scripts/stereo_marker_heatmap_template.R: marker/program heatmap template adapted from SpaSEG, human cortex, and GF/SPF cecum heatmap-style figure patterns.
When a plot requires a more specialized analysis output, use this skill together with the relevant analysis skill and still preserve the publication-figure requirements here.
Output Expectations
- Figure files and any table used for plotting.
- Plot variables, coordinate orientation, palette, and font settings.
- Any data filtering or aggregation applied only for display.
- Reused article code source, paper DOI, repository or code DOI, original file name, and dataset-specific edits.