What problem does it solve? Scientific figures made ad hoc end up visually inconsistent across a paper or thesis — different palettes, clustering choices, label handling, and axis conventions. This Skill encodes one lab's publication-figure defaults as reusable R and Python functions so every heatmap, scatter, stripchart, boxplot, and density plot follows the same conventions automatically. ## Core Features & Use Cases - pavlab_heatmap: Wraps pheatmap (R) and seaborn.heatmap (Python) with expression/correlation/raw modes, row Z-scoring with ±3 clipping, a black-centered divergent palette, grey NA cells, and automatic label hiding for large matrices. - pavlab_scatter, pavlab_stripchart, pavlab_boxplot, pavlab_density: Distributional and two-variable plots with N-adaptive rendering (solid points → alpha blending → hexbin), log2/log10 axis handling, swarm or jitter point layouts, and optional grand mean/median reference lines. - Shared palettes and style rules: Byte-identical divergent and black-body sequential palettes in both languages, plus cross-plot rules (no gridlines, outward ticks, SD-or-CI95 error bars, never SEM). - Use Case: A researcher needs a sample–sample correlation heatmap for a paper figure; calling pavlab_heatmap(cor_df, mode="correlation") produces a diagonal-masked, correctly-paletted PDF with no manual tuning. ## Quick Start Ask the AI to create a publication-quality heatmap of your expression matrix using the plotting skill's pavlab_heatmap function with default lab styling.