Molecular AOP Analyser — Key Event enrichment analysis
Molecular AOP Analyser
Upload a gene expression dataset (CSV, TSV, or TXT) to analyse it against molecular Adverse Outcome Pathways. The tool will guide you through choosing the right columns, visualising gene expression with a volcano plot, and running KE enrichment against a selected AOP.
New here? Try a curated demo dataset first to see how the tool works.
Upload Your Own Data
Experiment Information (Optional)
Providing experiment details helps generate comprehensive reports and enables better data organization.
Batch Analysis
Upload 2–10 gene expression files, tag each with condition metadata, and run enrichment analysis with a harmonised gene background across all files.
1. Upload Files
Upload 2–10 gene expression files (CSV, TSV, or TXT). All files must share the same column layout.
Drag files here or click to select
CSV, TSV, TXT — max 10 files, 10 MB each
2. Tag Conditions
Assign a condition label to each file. Dose and timepoint are optional but help with downstream comparison.
Specify the columns shared across all files. These will be detected automatically from the first file.
3. Analysis Settings
Batch Information
AOP Selection
Select the Adverse Outcome Pathway to use for all conditions in this batch.
Statistical Method
Applied equally to all conditions, so the comparison stays like-for-like. Fisher's exact tests the genes that pass the thresholds below. GSEA ranks every gene instead and reports a signed NES — which is what makes a coordinated shift visible across a dose or timepoint series — but the thresholds still define the per-condition significant-gene counts and the driver genes shown on the comparison page.
Enrichment Background
Fisher's exact test asks whether a gene set is over-represented among your significant genes relative to a background, so the background is part of the result, not a technicality. Two separate decisions define it: what counts as a gene in one condition, and how the conditions combine.
DESeq2 and limma withhold an adjusted p-value from low-count genes. Those genes were measured, they simply cannot be called significant. Excluding them from the background too selects against the genes an induced response is made of — low at baseline, strongly up on treatment — so a gene set built from inducible genes is scored against a universe that left them out.
Intersection reads as the conservative choice and is not. Each condition's universe has already been shaped by that condition's own filtering, so intersecting collapses the batch onto whichever condition was filtered hardest — on a dose series the background can end up equal to a single dose, with the rest contributing nothing. Choosing none means enrichment values are not comparable between conditions, because the background differs between them.
Log2FC Threshold
Minimum absolute log2 fold change for a gene to be considered significant. Applied equally to all conditions.
P-value Threshold
Genes with a raw p-value below this threshold are flagged as significant. Fisher's exact tests on that flag; under GSEA it only drives the reported gene counts. Applied equally to all conditions. Default 0.05.
Gene Set Resources
Curated KE gene-set resources to enrich against. Genes are pooled across the selected resources. Applied equally to all conditions.
Minimum Mapping Confidence
KE→pathway mappings are curated as High, Medium or Low confidence. Raising the threshold drops weaker mappings from each Key Event's gene set; a KE whose mappings are all below the threshold is not tested. Mappings without a recorded confidence are always kept. Applied equally to all conditions.
Applies to all three resources. WikiPathways mappings are filtered here; the Gene Ontology and Reactome gene sets are filtered by the Builder when they are fetched. A resource served from the reference files bundled with this image is used in full — the batch records which resources the threshold was actually applied to.