本文へ移動
cccskills
無料GitHub で公開

scientific-schematics

Designing scientific schematics, diagrams, and graphical abstracts. Covers tool selection (BioRender, Inkscape, Affinity, PowerPoint), design principles for pathway diagrams, mechanism schematics, experimental workflows, and journal graphical abstracts. Includes composition, icon sourcing, color for biological entities, and accessibility. Use when creating illustrative (not data-driven) scientific figures.

インストール方法を見る

含まれるファイル(1)

  • SKILL.md15.1 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

Scientific Schematics

Overview

Scientific schematics are illustrative figures that communicate biological mechanisms, experimental designs, or conceptual frameworks — as distinct from data-driven figures (graphs, heatmaps). A well-designed schematic makes a manuscript's key concept immediately comprehensible to a broad scientific audience, while a poorly designed one confuses reviewers and reduces the paper's impact. This guide covers the design decisions, tool selection, composition principles, and accessibility standards specific to biological schematics in peer-reviewed publications.

Key Concepts

1. Schematic Types and Their Purpose

Not all schematics serve the same function. Choosing the wrong type leads to overloaded or underspecified figures.

TypePurposeKey ElementsExamples
Mechanism diagramShow step-by-step molecular eventsProteins, DNA, membranes, arrows indicating causalityCRISPR cleavage, signaling cascade, transcription factor binding
Pathway diagramShow relationships in a networkNodes (proteins/metabolites), edges (activation/inhibition), directionMAPK signaling, metabolic flux, gene regulatory network
Experimental workflowShow experimental protocol as a figureNumbered steps, icons for equipment/samples, time arrowsSingle-cell sequencing pipeline, drug treatment protocol
Graphical abstract1-panel summary of entire paper2–4 key findings, minimal text, journal-specified dimensionsCell, Nature, PNAS graphical abstracts
Structural diagramShow molecular structure/conformationRibbon structure, surface representation, active siteProtein domain schematic, ligand binding pocket

Decision rule: If the reader needs to understand what happened step-by-step, use a workflow. If they need to understand why (cause and effect), use a mechanism/pathway diagram. If they need a 30-second paper summary, use a graphical abstract.

2. Tool Comparison

ToolBest ForLearning CurveCostFile Output
BioRenderQuick biological schematics; built-in icon libraryLowSubscription (~$99/month academic; free for draft)PNG, SVG, PDF (license required for publication)
InkscapeFull vector editing; open-sourceMedium–HighFreeSVG, PDF, PNG, EMF
Affinity DesignerProfessional vector + raster hybridMedium~$55 one-timeSVG, PDF, AFDESIGN
Adobe IllustratorIndustry standard; full featuresHigh~$55/monthSVG, PDF, AI, EPS
PowerPoint / KeynoteQuick drafts; non-scientistsLowIncluded in MS/ApplePNG (low resolution), PDF
draw.io / LucidchartFlowcharts, pathway diagramsLowFree/subscriptionSVG, PNG, XML
ChimeraX / PyMOLMolecular structure rendersMediumFree (academic)PNG, TIFF, movie

Key difference: BioRender provides curated biological icons (cells, proteins, instruments) that are publication-quality and royalty-free under the publication license. For complex custom graphics, Inkscape (free) or Affinity Designer (paid) offer more flexibility.

3. Color Usage in Biological Schematics

Color in biological schematics carries semantic meaning and must be used consistently.

Conventional biological color assignments:

  • DNA: double helix colored in blue/navy or gray
  • mRNA: orange or yellow
  • Protein: varying; use consistent color per protein family across all figures in a paper
  • Cell nucleus: light gray or blue fill
  • Cell membrane: tan/beige bilayer
  • Mitochondria: orange
  • Lysosomes: red/purple

Signal direction conventions:

  • Arrows: activation (→ solid arrowhead), inhibition (⊣ flat bar), indirect (- - →)
  • Color coding arrows: green = activating, red = inhibiting, gray = neutral/context

Accessibility:

  • Never use red–green encoding alone; add shape cues (✓/✗, solid/dashed)
  • Test with CVD simulator before submission
  • Use at least 3:1 contrast ratio for text on colored backgrounds

4. Composition and Layout Principles

Visual hierarchy: The most important element should be the largest and most central. Support elements (labels, arrows, scale bars) should be visually subordinate.

Element spacing: Maintain consistent padding between grouped elements. Crowded diagrams cause readers to misread association between elements.

Arrow semantics: Every arrow must convey exactly one relationship. Do not use arrows interchangeably for "leads to," "causes," "physically binds," and "is required for" — use distinct arrow styles for each.

Text minimalism: Labels should be noun phrases only (not full sentences). Move explanatory text to the figure caption, not into the diagram itself.

Decision Framework

What is this schematic trying to communicate?
│
├── A sequence of experimental steps
│   └── → Workflow diagram
│       ├── Numbered panels or numbered arrows
│       ├── Icons for equipment (instrument, pipette, plate, animal)
│       └── Time axis or step axis
│
├── How a molecule works mechanistically
│   └── → Mechanism diagram
│       ├── Protein shapes (surface or cartoon)
│       ├── Conformational change arrows
│       └── Before/after state panels
│
├── How multiple components interact in a network
│   └── → Pathway diagram
│       ├── Nodes = proteins/metabolites (circles, rectangles)
│       ├── Edges = arrows (activation = →, inhibition = ⊣)
│       └── Compartment boxes (nucleus, cytoplasm, extracellular)
│
├── A single-panel paper summary
│   └── → Graphical abstract
│       ├── Check journal specifications (size, text policy)
│       ├── Show: question → approach → key finding → implication
│       └── Minimal text; max 3–4 panels
│
└── A 3D molecular structure
    └── → Structural diagram
        ├── Use ChimeraX, PyMOL, or RCSB PDB viewer for render
        ├── Annotate active site/binding pocket
        └── Add scale bar (Å or nm)
ScenarioToolFormatKey Consideration
First-author manuscript, fast turnaroundBioRenderPDF/SVGRequires publication license for final submission
Lab with ongoing schematic needsInkscape + master templateSVGReusable elements; free; no vendor lock-in
Graphical abstract for Cell/NatureBioRender or IllustratorPNG at journal specCheck exact pixel dimensions and text policy
Molecular mechanism with protein structureChimeraX + InkscapePNG render + SVG annotationUse PDB structure; annotate key residues
Pathway diagram with many nodesdraw.io / CytoscapeSVG/PDFAuto-layout for networks with >20 nodes

Best Practices

  1. Start with a sketch before opening any tool: Draw the schematic on paper first. Determine what elements are needed, their hierarchy, and the spatial relationships. Opening software first leads to layout decisions driven by tool constraints, not by communication goals.

  2. Define a consistent visual vocabulary before drawing the first panel: Assign colors, shapes, and arrow styles to each biological entity type at the project level. Document these assignments in a legend or lab style guide. Inconsistency across figures in a paper signals carelessness and confuses reviewers.

  3. Use the journal's graphical abstract template: Every high-impact journal (Cell, Nature, PNAS, PLOS) publishes exact dimensions, resolution requirements, and font rules for graphical abstracts. Download the template before starting — resizing after the fact destroys aspect ratios and makes text too small.

  4. Export at publication resolution from the vector source: Always export final figures from the original vector file (SVG, AI, AFDESIGN) at the target DPI, never by screenshot or screen capture. Screenshots are raster at screen resolution (~72 dpi); journals require 300–600 dpi.

  5. Label arrows and connections in captions, not in the diagram: Arrow labels ("activates," "phosphorylates," "recruits") add clutter and reduce diagram readability. Move these relationships to the figure caption using panel reference labels (A, B, C). The diagram should be self-explanatory to someone who reads the caption.

  6. Group related elements visually using proximity, enclosure, and color: Elements that belong together conceptually should be spatially close, share a background color, or be enclosed in a bounding box. Do not use arbitrary spatial arrangements — viewers interpret proximity as association.

  7. Version-control your source files: Store original editable source files (.svg, .afdesign, .ai, .brd) in version-controlled storage alongside the manuscript. Journals frequently request revised figures during revision, and re-creating from a PNG is extremely time-consuming.

Common Pitfalls

  1. Using inconsistent arrow styles across a figure

    • What goes wrong: Solid arrows and dashed arrows mixed arbitrarily make readers unsure whether dashed means "indirect," "hypothetical," or merely a stylistic choice.
    • How to avoid: Define an arrow legend: solid arrow = direct activation, dashed = indirect, flat bar = inhibition. Use this consistently in every panel of every figure in the paper.
  2. Overloading a single panel with too many elements

    • What goes wrong: Pathway diagrams with 20+ nodes and 30+ arrows cannot be read in a journal figure at 170 mm width. Reviewers ask for simplification; revisions are time-consuming.
    • How to avoid: Limit mechanism panels to ≤8 elements. Move complex pathway context to supplementary figures. Focus the main-text schematic on the core novel finding.
  3. Using BioRender without purchasing the publication license

    • What goes wrong: BioRender figures created under the free or draft license cannot be published. Submitting without proper licensing violates BioRender's terms of service and risks manuscript rejection.
    • How to avoid: Purchase the publication license before creating the final figure version. BioRender generates a citation string that must be included in the figure legend.
  4. Text inside the schematic at 6 pt or smaller at print size

    • What goes wrong: Labels inside pathway nodes or protein icons become illegible at publication dimensions (85 mm single-column). Reviewers flag this; revision requires recreating the entire figure.
    • How to avoid: Work at print size from the start. In Inkscape: set canvas to 85 mm × target height. Minimum label size: 7 pt at 100% print scale.
  5. Red–green arrow encoding for activation–inhibition without shape cues

    • What goes wrong: ~8% of male readers have red–green color vision deficiency and cannot distinguish activation (green) from inhibition (red) arrows.
    • How to avoid: Use distinct arrow shapes: → for activation, ⊣ for inhibition. Color is redundant encoding, not the only cue.
  6. Copying schematic elements from published papers without redrawing

    • What goes wrong: Reproducing figures from published papers without permission violates copyright. This applies to diagrams and icons, not just photographs.
    • How to avoid: Use BioRender's licensed icon library, Reactome's freely licensed pathway diagrams (CC-BY), or Bioicons (free, open icons). Always redraw; never screenshot.
  7. Missing scale bars in structural diagrams

    • What goes wrong: Protein structure renderings without scale bars prevent readers from assessing relative sizes of domains, ligands, or interfaces.
    • How to avoid: Add a scale bar (e.g., "10 Å" or "2 nm") using ChimeraX's scalebar command or annotate in Inkscape after export.

Workflow

  1. Define the message — write one sentence stating what the schematic must communicate
  2. Sketch on paper — rough layout, element list, arrow types
  3. Choose tool — BioRender for icon-heavy biology; Inkscape for custom vectors; ChimeraX for structures
  4. Set canvas to print dimensions — match journal column width before placing any element
  5. Build in layers — background → structural elements (membranes, organelles) → protein/molecule icons → arrows → labels
  6. Apply visual vocabulary — consistent colors, arrow styles, fonts per lab style guide
  7. Run accessibility check — CVD simulation; test grayscale version
  8. Export from vector source — PDF or TIFF at target DPI (≥300 dpi raster, vector for line art)
  9. Archive source file — commit .svg/.afdesign/.brd to version control alongside manuscript

Protocol Guidelines

  1. BioRender publication workflow: Create figure → File → Publish Figure → Generate Attribution → copy citation string into figure legend. The attribution reads: "Created with BioRender.com" with a unique agreement ID. Without this, the figure cannot be legally published.
  2. Inkscape CMYK-safe workflow: Inkscape does not natively support CMYK. For CMYK journals: design in RGB → export to PDF → convert using Ghostscript (gs -sDEVICE=pdfwrite -dUseCIEColor). Verify colors after conversion.
  3. ChimeraX protein structure render: open 6YYT → graphics silhouettes true → lighting soft → saveimage figure.png width 2400 height 2400 → annotate in Inkscape.
  4. Multi-panel figure assembly in Inkscape: Use Inkscape extensions → Grids → set column guides at journal column widths. Import each panel as a linked SVG to enable per-panel updates without recreating the composite.

Further Reading

Related Skills

  • scientific-visualization — data-driven figures (graphs, heatmaps) as distinct from illustrative schematics
  • scientific-manuscript-writing — figure legend writing and integration of schematics into papers
  • latex-research-posters — assembling schematics into poster layouts

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

API + Python SDK for ordering cell-free protein expression and binding assays. Submit sequences for expression (10–100 µg), measure binding affinity (KD) against targets, track status, and retrieve results programmatically — no wet-lab setup. Built for ML-guided directed evolution and antibody/nanobody optimization. Requires Adaptyv account and API key.

日本語の概要は準備中です。原文の説明を表示しています。

jaechang-hits/SciAgent-Skills3762026年9月29日 更新

aeon

無料

scikit-learn compatible Python toolkit for time series ML: classify, cluster, regress, segment, transform with 30+ algorithms (ROCKET, InceptionTime, KNN-DTW, HIVE-COTE, WEASEL). Handles panel, multivariate, and unequal-length series. Maintained successor to sktime. Alternatives: sktime (larger ecosystem), tslearn (fewer algorithms), catch22 (features only).

日本語の概要は準備中です。原文の説明を表示しています。

jaechang-hits/SciAgent-Skills3762026年9月29日 更新

AiZynthFinder retrosynthetic route planning (CASP) from AstraZeneca Molecular AI. Monte Carlo tree search guided by a template-based neural expansion policy recursively disconnects a target SMILES until precursors are found in a purchasable stock. Covers config.yml (v4 format), aizynthcli batch screening, the AiZynthFinder/AiZynthExpander Python API, one-step disconnections, custom stocks via smiles2stock, scorers, Retro*/breadth-first/DFPN search alternatives, and reading output.json.gz / trees.json. Use for synthesis route planning, synthesizability screening, and building-block/precursor search. For reaction barriers use neb-irc-activation-energy; for 2D reaction scheme drawing use rdkit-chemdraw-cdxml.

日本語の概要は準備中です。原文の説明を表示しています。

jaechang-hits/SciAgent-Skills3762026年9月29日 更新

Access AlphaFold DB's 200M+ predicted structures by UniProt ID. Download PDB/mmCIF, analyze pLDDT/PAE, bulk-fetch proteomes via Google Cloud. For experimental structures use PDB; for prediction use ColabFold or ESMFold.

日本語の概要は準備中です。原文の説明を表示しています。

jaechang-hits/SciAgent-Skills3762026年9月29日 更新

Annotated matrices for single-cell genomics. Stores X with obs/var metadata, layers, embeddings (obsm/varm), graphs (obsp/varp), uns. Use for .h5ad/.zarr I/O, concatenation, scverse integration. For analysis use scanpy; for probabilistic models use scvi-tools.

日本語の概要は準備中です。原文の説明を表示しています。

jaechang-hits/SciAgent-Skills3762026年9月29日 更新

GRN inference from expression via GRNBoost2 (gradient boosting) or GENIE3 (Random Forest). Load matrix, filter by TFs, infer TF-target-importance links, save network. Dask-parallelized to single-cell scale. Core SCENIC component.

日本語の概要は準備中です。原文の説明を表示しています。

jaechang-hits/SciAgent-Skills3762026年9月29日 更新

jaechang-hits のスキルをすべて見る

このスキルの問題を報告する