Free AI-Assisted

Matched Null Gene Set Generator

Generate confound-matched random null gene sets for enrichment negative controls with SMD balance metrics, reproducible seeds, and one-click GMT export. Built-in AI agent assistant support.

Open tool workspace

Need help with bioinfo analysis? See our bioinformatics services.

Key facts

Key facts about Matched Null Gene Set Generator
FactValue
Query limit500 genes per run
OrganismsHuman, Mouse, Rat, Zebrafish, Fly, C. elegans, S. cerevisiae
MatchingExpression, length, GC%, biotype; 3/5/10 bins
Balance metricsSMD and KS p-value per covariate
ExportGMT, CSV, Methods text, R/Python snippets
Runs in browserYes — client-side; optional API/MCP
AI assistantBuilt-in; configures matching via manifest

What it does

Enrichment tests need defensible negative controls. Naive random gene sets inherit expression, length, and GC biases that inflate false positives. Matched Null Gene Set Generator builds stratified null sets in your browser: paste a query list, choose matching covariates, and export GMT files ready for GSEA, fgsea, or clusterProfiler.

The Generator tab walks through the full workflow on one page. Paste up to 500 gene symbols or Ensembl IDs, select organism (Human, Mouse, Rat, Zebrafish, Fly, Worm, or Yeast), and enable matching on expression percentile, transcript length, GC content, and optionally biotype. Choose tercile, quintile, or decile binning; set the number of null sets (1–50) and an optional seed. Query genes are excluded from the background pool before sampling.

After generation, review balance badges showing Standardized Mean Difference (SMD) and Kolmogorov–Smirnov p-values per covariate. Export GMT or CSV, copy the first null set, or grab publication-ready Methods text plus R and Python snippets. Session history on the History tab recalls recent runs with seed, query size, and matching criteria for one-click reload.

Why researchers use it

  • Replace naive uniform sampling with expression-, length-, and GC-matched negative controls
  • Verify match quality with SMD < 0.1 balance badges before running permutation tests
  • Export GMT files directly compatible with GSEA, fgsea, and clusterProfiler
  • Generate 1–50 null sets in one click with reproducible seeds for methods sections
  • Use default protein-coding backgrounds per organism or upload a custom expressed-gene universe
  • Exclude query genes automatically so null sets never overlap the input list

Best for

  • Building matched negative controls for GSEA or fgsea permutation workflows
  • Generating null gene sets that control for expression and gene-length bias
  • Batch null-set generation for robust enrichment benchmarking
  • Preparing GMT files for clusterProfiler custom background tests
  • Documenting null-set parameters with auto-generated Methods text

When to use this vs alternatives

Use Matched Null Gene Set Generator when you need covariate-matched null sets for enrichment—not just list overlap statistics. For comparing multiple gene lists with Venn diagrams and hypergeometric p-values, use Smart List Overlap. For merging annotation tables before enrichment, use Bio Table Matchmaker.

What makes it different

Molbiotools and similar utilities sample uniformly from static genome lists with no covariate matching. Seurat control-gene logic is powerful but opaque and tied to single-cell workflows. Matched Null Gene Set Generator is a standalone, browser-based tool focused on enrichment null models: stratified sampling with adaptive bin merging, quantified balance metrics, and direct GMT export—without installing R packages or writing one-off Perl scripts.

Unlike server pipelines, all computation runs client-side after annotation data loads. Your gene lists never leave the browser unless you use the optional API.

How to get started

  1. Open the Generator tab and click Load Example for a 30-gene human test set.
  2. Confirm matching options (expression, length, GC on by default) and bin resolution.
  3. Set null set count and optional seed, then click Generate null sets.
  4. Review balance badges — aim for SMD < 0.1 on each matched covariate.
  5. Export GMT for fgsea or copy Methods text for your paper.
  6. Switch to History to reload a previous run configuration.

Frequently asked questions

What background universe is used?
By default, all protein-coding genes for the selected organism with expression, length, and GC annotation from GTEx (human) or Expression Atlas RNA-seq (other species). You can upload a custom CSV/TSV background with symbol, Ensembl, expression, length, and GC columns.
What does adaptive binning mean?
If a query gene falls in a bin with too few background candidates, adjacent bins are merged until enough genes are available. A warning shows how many draws used merged bins.
Are query genes included in null sets?
No. All query genes are removed from the background pool before sampling. Null sets never overlap your input list.
How is balance evaluated?
For each continuous covariate (expression, length, GC), the tool reports Standardized Mean Difference (SMD) and a two-sample Kolmogorov–Smirnov p-value. SMD < 0.1 is labeled Optimal; 0.1–0.2 is Acceptable.
Can I use an AI agent or MCP?
Yes. The workspace assistant can configure matching from natural language. External agents call the same handler via [API & MCP](/tools/developers) using pepkio_matched-null-gene-set-generator.

Client source code & registry

Open tool workspaceBack to tools

Last updated . Pepkio builds free lab calculators alongside bioinformatics CRO services.