pathXcite: Literature-Based Gene Enrichment

What pathXcite does

pathXcite is a standalone application for literature-based functional genomics. Instead of requiring expression data, it mines published articles to extract, rank, and analyze genes relevant to your topic: and performs enrichment against built-in or custom gene set libraries.

• Retrieve literature: query PubMed or PMC, review results, and build a tailored corpus.
• Extract gene mentions: automatically identify and standardize genes using PubTator Central.
• Rank by relevance: use simple frequency or GF-IDF (Gene-Frequency-Inverse-Document-Frequency) to emphasize topic-specific genes.
• Run enrichment: test against hundreds of curated libraries or your own .gmt files.
• Visualize & export: interactively explore enriched terms and export plots and tables.

All data remain local in self-contained project folders, enabling reproducibility, portability, and offline analysis once literature is cached.

First analysis in five steps

1. Create a project: Start a new project; a dedicated folder stores corpus, annotations, and results.
2. Add articles: Paste PMIDs/PMCIDs or use the built-in browser to search and select papers.
3. Extract genes: Annotate articles to identify gene mentions and map them to NCBI Gene IDs.
4. Rank and filter: Choose absolute frequency or GF-IDF to rank genes by prominence or specificity.
5. Enrich: Select one or more gene set libraries, from pathways and phenotypes to drug responses, and run enrichment.

See the Quick Start guide for a full walkthrough.

Gene ranking strategies

Absolute Frequency

Most mentioned genes

Counts how many times a gene appears in your corpus. Highlights widely studied, canonical genes.

Intuitive and straightforward Captures well-established core genes Can overrepresent highly cited genes

GF-IDF

Topic-specific genes

Adjusts for how often each gene appears across the entire literature, emphasizing those unusually enriched in your corpus.

Highlights context-specific genes Reduces literature popularity bias May downrank broadly studied genes

Running enrichment with both strategies often reveals complementary patterns: canonical and topic-specific gene landscapes.

Broad support for gene set libraries

pathXcite supports enrichment against a wide range of curated libraries. More than 240 are included, and additional libraries can be loaded via .gmt files.

• Biological processes & functions: Gene Ontology (BP, MF, CC), HPO
• Pathways: KEGG, Reactome, WikiPathways, BioCarta, BioPlanet
• Regulation: ChEA, ENCODE TF targets, JASPAR motifs, TRRUST
• Disease & phenotype: DisGeNET, OMIM, ClinVar, Orphanet, MGI, KOMP2
• Drug & chemical signatures: LINCS L1000, DSigDB, GEO perturbation libraries
• Cell types & tissues: GTEx, ARCHS4, CellMarker, Azimuth, Allen Brain Atlas
• Complexes & interactions: CORUM, BioPlex, virus-host PPIs
• Custom libraries: Import your own .gmt files.

See the tutorials on Gene Set Libraries for details.

Inputs and outputs

Analysis tips

• Use an NCBI API key to speed up literature retrieval.
• Curate your corpus: focus on topic-relevant articles for higher-quality annotations.
• Match libraries to your question: regulatory libraries for transcriptional hypotheses, drug libraries for repurposing, disease libraries for clinical context.
• Compare rankings: run enrichment on both frequency and GF-IDF lists to capture complementary signals.
• Use project folders: analyses are fully reproducible and can be reopened at any time.

System requirements