Volcano Plot

A volcano plot displays the relationship between statistical significance and effect size for a large number of features — typically genes, proteins, or other molecular entities — tested simultaneously. It is a standard figure in differential expression analysis and similar high-dimensional comparisons.

What the figure shows

The axes

The x-axis shows the fold change (or log₂ fold change) between two conditions. Positive values indicate higher expression or abundance in one condition; negative values indicate lower.

The y-axis shows −log₁₀ of the p-value. This transformation makes small p-values appear high on the plot. A value of 1.3 on this axis corresponds to p = 0.05.

Color coding

Points are colored by their position relative to the significance and fold change thresholds:

• Red — statistically significant and exceeds the fold change threshold (up or down)
• Orange — statistically significant but does not exceed the fold change threshold
• Gray — not statistically significant

Threshold lines

A horizontal dashed line marks the p-value threshold (default: p = 0.05, corresponding to y = 1.3). Two vertical dashed lines mark the fold change thresholds (default: ±1.0).

Default thresholds

To use different thresholds, specify them in the Chat:

• "Use a fold change threshold of 1.5"
• "Set the significance threshold to 0.01"

These defaults are display thresholds, not universal biological cutoffs. They are meant to provide a readable starting view for ranking features, but different assays and domains often need different fold-change and significance criteria.

Required columns

A volcano plot requires two numeric columns:

• A fold change column — values representing the magnitude of change between conditions (log₂ FC is standard but any numeric scale works)
• A p-value column — values must be greater than zero; rows with p ≤ 0 are excluded

Licklider detects these columns automatically from common naming patterns such as log2fc, logfc, fold_change, pval, and padj.

Licklider cannot automatically determine whether the imported p-value column is unadjusted or already corrected for multiple testing unless that meaning is made clear in the selected column. It also cannot guarantee that the fold-change values come from the preprocessing, normalization, or model that best matches your assay.

Those limits matter because a volcano plot can look highly convincing even when the underlying p-values are not multiplicity-adjusted or when the fold-change scale is not the one your readers expect.

What is not currently shown

Gene or feature labels on individual points are not currently displayed. To identify specific points of interest, inspect the underlying data table.

This is intentional. Dense high-dimensional results can become unreadable if many labels are shown at once, so the current public surface keeps the plot focused on global ranking structure rather than point-by-point annotation.

Design rationale and references

Licklider treats volcano plot as a descriptive ranking surface built from imported statistics, not as a full analysis engine. That is why the page emphasizes what columns must already exist and keeps the statistical test generation outside the scope of this figure.

The page also recommends adjusted p-values when available because volcano plots are most often used in settings with many simultaneous feature tests. Without multiplicity-aware p-values, the upper part of the plot can make noise look like a set of robust discoveries.

The default thresholds are included to give users a legible starting view, not to assert that the same cutoff is scientifically appropriate for every omics workflow. In practice, the meaning of a "large enough" fold change depends on assay noise, normalization, and the biological context.

References

1. Cui, X., & Churchill, G. A. (2003). Statistical tests for differential expression in cDNA microarray experiments. Genome Biology, 4(4), 210. https://doi.org/10.1186/gb-2003-4-4-210
2. Love, M. I., Huber, W., & Anders, S. (2014). Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biology, 15, 550. https://doi.org/10.1186/s13059-014-0550-8

What this page does not cover

• The statistical tests used to produce fold change and p-values — these are computed outside Licklider and imported as columns
• Multiple testing correction — if adjusted p-values are available, use those as the p-value column