Volcano Plot

A volcano plot displays the fold-change (x-axis) versus the statistical significance (y-axis, typically 竏値og10(adjusted p-value)) of many features simultaneously. It is the standard visualization for differential expression, proteomics, and metabolomics analyses. The critical issue: the p-values shown must be multiple-testing-corrected —a volcano plot showing unadjusted p-values for thousands of features is statistically invalid.

STEP 1 —When to Use

Use a volcano plot when: you have tested many features simultaneously (gene expression, proteins, metabolites) and want to show both statistical significance and effect size (fold-change) in one view; you want to identify features that are both statistically significant AND have a large fold-change; the data come from a high-throughput comparison (n features typically >= 100).

Multiple testing is always present in volcano plots. Every volcano plot involves thousands of simultaneous tests. Unadjusted p-values on the y-axis are never acceptable for a volcano plot —the expected number of false positives at unadjusted p < 0.05 across 10,000 features is 500.

[Image placeholder: Volcano plot with log2(fold-change) on x-axis and 竏値og10(adjusted p-value, BH-FDR) on y-axis. Points colored: grey (not significant), blue (significant by p only), red (significant by both p and FC threshold). Threshold lines: vertical dashed at ﾂｱ1 (2-fold change), horizontal dashed at 竏値og10(0.05). Feature labels for top 10 most significant shown with repel labels to avoid overlap.]

Volcano plot: fold-change vs. BH-FDR-adjusted p-value. Threshold lines for FC (ﾂｱ1 log2FC) and significance (BH-FDR q 竕､ 0.05) shown as dashed lines.

STEP 2 —Journal Requirement

Multiple testing correction method must be stated: BH-FDR (Benjamini-Hochberg, default for most), FWER (Bonferroni, for strong control), or other.
The adjusted p-value threshold used must be stated (typically q 竕､ 0.05 for FDR).
The fold-change threshold used must be stated and justified.
n features tested in total, n features significant, must be stated.
The legend or methods must identify up-regulated and down-regulated features by the color coding.

STEP 3 —Licklider Implementation

Input

Feature names, fold-change values, and p-values (unadjusted)
Multiple testing correction method: BH-FDR (default), Bonferroni, Holm, or pre-adjusted input
Significance threshold (q 竕､ 0.05 default) and fold-change threshold (FC >= 2, log2FC >= 1 default)
Label top N features (by adjusted p-value, n=10 default)

Output

Scatter: log2(FC) on x-axis, 竏値og10(adjusted p-value) on y-axis
Color coding: significant + above FC threshold (red/blue for up/down), significant only (light color), non-significant (grey)
Threshold lines: vertical at ﾂｱFC threshold, horizontal at significance threshold
Feature labels for top N with non-overlapping label placement
Legend with color scheme, thresholds, and correction method
Summary count: "N up-regulated, N down-regulated (q 竕､ 0.05, |log2FC| >= 1)"

[Image placeholder: Licklider volcano plot. 8,000 features total. 127 significant (BH-FDR q 竕､ 0.05 and |log2FC| >= 1): 74 up-regulated (red, right of FC threshold), 53 down-regulated (blue, left). Top 10 significant features labeled. Threshold lines shown. Legend: "BH-FDR corrected (q 竕､ 0.05), |log2FC| >= 1. 8,000 features tested."]

Licklider volcano plot: BH-FDR correction labeled, feature counts by direction, top features labeled, threshold lines, summary count in legend.

STEP 4 —Draft Output (Draft / Needs review)

Volcano plot displaying log2 fold-change (x-axis) versus 竏値og10 adjusted
p-value (y-axis) for 8,000 features. Statistical significance was assessed
using [test], and p-values were adjusted for multiple comparisons using
the Benjamini-Hochberg false discovery rate (FDR) method. Thresholds:
FDR q 竕､ 0.05 (horizontal dashed line) and |log2FC| >= 1 (vertical dashed
lines). 127 features met both criteria: 74 up-regulated (red) and 53
down-regulated (blue). Top 10 features by adjusted p-value are labeled.