SuperAPI AI Directory

scientific-publication-figures

General & other

View repository

Updated 10-15 21:03

About this skill

Scientific Publication-Quality Figure Generation

A skill for creating scientific figures at Nature/Science/Cell publication level with matplotlib/seaborn. It provides the 15 figure templates established in Exp-10 and the style settings common to all experiments.

When to Use

  • When creating high-quality figures for journal submission
  • When composing a multi-panel composite figure
  • When DPI 300 or higher and appropriate font sizes and color palettes are required
  • When figures need to comply with a journal's style guide

Quick Start

Global Style Settings

Always apply the following settings at the beginning of the script.

import matplotlib.pyplot as plt
import matplotlib as mpl

def setup_publication_style():
    """Nature/Science 品質の matplotlib rcParams を設定する。"""
    plt.rcParams.update({
        # フォント
        "font.family": "sans-serif",
        "font.sans-serif": ["Arial", "Helvetica", "DejaVu Sans"],
        "font.size": 10,
        "axes.titlesize": 12,
        "axes.labelsize": 11,
        "xtick.labelsize": 9,
        "ytick.labelsize": 9,
        "legend.fontsize": 9,

        # 線・マーカー
        "axes.linewidth": 1.2,
        "lines.linewidth": 1.5,
        "lines.markersize": 6,

        # スパイン
        "axes.spines.top": False,
        "axes.spines.right": False,

        # グリッド
        "axes.grid": False,

        # 保存設定
        "savefig.dpi": 300,
        "savefig.bbox": "tight",
        "savefig.transparent": False,

        # Figure 背景
        "figure.facecolor": "white",
        "axes.facecolor": "white",
    })

Color palette

# Nature-inspired パレット(Exp-10)
NATURE_PALETTE = [
    "#E64B35",  # 赤
    "#4DBBD5",  # 水色
    "#00A087",  # 緑
    "#3C5488",  # 紺
    "#F39B7F",  # サーモン
    "#8491B4",  # グレー青
    "#91D1C2",  # ミント
    "#DC9C6D",  # オレンジ
]

# 材料・群カテゴリ用
MATERIAL_PALETTE = {
    "Set2": "seaborn Set2(最大 8 色、差別化しやすい)",
    "tab10": "matplotlib tab10(最大 10 色)",
    "Paired": "seaborn Paired(最大 12 色、対比較向き)",
}

# 連続値用
CONTINUOUS_CMAPS = {
    "viridis": "知覚均一(デフォルト推奨)",
    "RdBu_r": "相関ヒートマップ(中心=0)",
    "YlOrRd": "単方向の強度(密度、温度)",
    "coolwarm": "二方向(応力の引張 vs 圧縮)",
}

Common Save Function

def save_fig(fig, filename, dpi=300, formats=("png",)):
    """図を指定された形式で保存する。"""
    for fmt in formats:
        filepath = f"figures/{filename}.{fmt}"
        fig.savefig(filepath, dpi=dpi, bbox_inches="tight",
                    facecolor="white", edgecolor="none")
    plt.close(fig)

Main Figure Templates

1. Violin + Strip Plot (with significance markers)

import seaborn as sns
from scipy import stats

def plot_violin_with_significance(df, x, y, pairs=None, figsize=(10, 6)):
    fig, ax = plt.subplots(figsize=figsize)
    sns.violinplot(data=df, x=x, y=y, palette=NATURE_PALETTE,
                   inner=None, alpha=0.3, ax=ax)
    sns.stripplot(data=df, x=x, y=y, palette=NATURE_PALETTE,
                  size=3, alpha=0.5, ax=ax)

    # 有意差ブラケット(pairs 指定時)
    if pairs:
        y_max = df[y].max()
        for i, (g1, g2) in enumerate(pairs):
            d1 = df[df[x] == g1][y]
            d2 = df[df[x] == g2][y]
            stat, pval = stats.mannwhitneyu(d1, d2)
            stars = "***" if pval < 0.001 else "**" if pval < 0.01 else "*" if pval < 0.05 else "ns"
            h = y_max * (1.05 + i * 0.08)
            # ブラケット描画は adjustText か手動 annotate
            ax.text((list(df[x].unique()).index(g1) +
                    list(df[x].unique()).index(g2)) / 2,
                   h, stars, ha="center", fontsize=12, fontweight="bold")
    save_fig(fig, f"violin_{y}")

2. Forest Plot (meta-analysis style)

def plot_forest(effects, ci_lower, ci_upper, labels, figsize=(10, 8)):
    fig, ax = plt.subplots(figsize=figsize)
    y_pos = range(len(labels))
    ax.errorbar(effects, y_pos, xerr=[np.array(effects) - np.array(ci_lower),
                np.array(ci_upper) - np.array(effects)],
                fmt="D", color=NATURE_PALETTE[0], markersize=8,
                capsize=4, linewidth=2)
    ax.axvline(0, color="gray", linestyle="--", linewidth=1)
    ax.set_yticks(y_pos)
    ax.set_yticklabels(labels)
    ax.set_xlabel("Effect Size")
    ax.set_title("Forest Plot", fontweight="bold")
    save_fig(fig, "forest_plot")

3. Multi-panel Composite Figure

import matplotlib.gridspec as gridspec

def create_composite_figure(plot_functions, layout=(2, 3),
                            figsize=(18, 12), panel_labels=True):
    """
    複数の描画関数を composite figure にまとめる。
    plot_functions: [(func, kwargs), ...] のリスト
    """
    fig = plt.figure(figsize=figsize)
    gs = gridspec.GridSpec(*layout, figure=fig, hspace=0.35, wspace=0.3)

    for i, (func, kwargs) in enumerate(plot_functions):
        row, col = divmod(i, layout[1])
        ax = fig.add_subplot(gs[row, col])
        func(ax=ax, **kwargs)

        if panel_labels:
            ax.text(-0.1, 1.1, chr(65 + i),  # A, B, C, ...
                   transform=ax.transAxes, fontsize=16,
                   fontweight="bold", va="top")

    save_fig(fig, "composite_figure")

Figure Size Guidelines

Purpose Size (inches) Notes
1 column (3.5, 3.0) Nature 1 column width ≈ 89 mm
1.5 columns (5.5, 4.0) Intermediate size
2 columns (7.0, 5.0) Nature 2 column width ≈ 183 mm
Full page (7.0, 9.0) A4 height excluding margins

ToolUniverse Integration

TU Key Tool name Integration details
biotools bio.tools Visualization tool registry search

References

Output Files

File Format
figures/*.png PNG (300 DPI)
figures/*.svg SVG (vector, for submission)
figures/*.pdf PDF (vector, for LaTeX)

Reference Experiments

  • Exp-10: Complete set of 15 figure templates
  • Exp-11〜13: Inheritance of rcParams settings and color palettes