Reactome Pathway Browser

Tier: 2 — Useful | Category: Pathway Knowledgebase / Enrichment Analysis

Reactome is a free, open-source, curated and peer-reviewed pathway knowledgebase covering human biology ^{1The Reactome Pathway Knowledgebase 20242024 · Nucleic Acids Research · DOI 10.1093/nar/gkad1025Open reference}. It currently catalogs over 2,600 human pathways and 14,000+ reactions spanning metabolism, signaling, gene expression, DNA repair, and immune function. The Reactome Pathway Browser provides an interactive visualization interface, and the ReactomePA R/Bioconductor package enables programmatic pathway enrichment analysis for transcriptomics and proteomics datasets ^{2ReactomePA: an R/Bioconductor package for reactome pathway analysis and visualization2010 · Molecular BioSystems · DOI 10.1093/bioinformatics/btq591 · PMID 20952405Open reference}.

Key Capabilities

• Over-representation analysis (ORA): Test whether gene sets from differential expression experiments are enriched in Reactome pathways

• Gene set enrichment analysis (GSEA): Ranked-list enrichment using Reactome pathway gene sets

• Pathway hierarchy browser: Navigate curated pathway hierarchy from top-level processes to molecular reactions

• Cross-species mapping: Human pathways mapped to model organisms via Ensembl Compara

• Integration: BioPAX export, API access, Cytoscape plugin (ReactomeFIViz)

Applications in Neurodegeneration Research

Reactome’s curated pathway hierarchy is particularly valuable for neurodegeneration research because it includes detailed mechanistic annotations for:

• Autophagy and mitophagy pathways (e.g., PINK1/Parkin, mTOR-dependent autophagy) central to Parkinson’s and ALS

• Unfolded protein response (UPR) and ER stress pathways activated in tauopathies

• Neuroinflammation signaling (NF-κB, MAPK, JAK-STAT) relevant to microglial activation in Alzheimer’s disease

• Tau protein binding partners and post-translational modification pathways

• Complement system and synaptic pruning pathways relevant to early AD neuronal loss

Tutorial: Pathway enrichment with ReactomePA

# Install required packages
if (!requireNamespace("ReactomePA", quietly = TRUE))
    BiocManager::install("ReactomePA")

library(ReactomePA)
library(org.Hs.eg.db)
library(AnnotationDbi)

# Example: AD GWAS hits from GEO dataset GSE5281
# Convert gene symbols to Entrez IDs
ad_genes <- c("APP", "PSEN1", "MAPT", "APOE", "TREM2", "BIN1", "CLU",
              "ABCA7", "CR1", "PICALM", "MS4A6A", "CD33", "EPHA1")

entrez_ids <- mapIds(org.Hs.eg.db, keys = ad_genes,
                     column = "ENTREZID", keytype = "SYMBOL")

# Run pathway enrichment
enrichment <- enrichPathway(
    gene = na.omit(entrez_ids),
    pvalueCutoff = 0.05,
    readable = TRUE
)

# View top pathways
head(as.data.frame(enrichment)[, c("Description", "GeneRatio", "p.adjust")])

# Visualize
dotplot(enrichment, showCategory = 15)

Reactome Pathways Relevant to Neurodegeneration

Pathway	Reactome ID	Relevance
Autophagy	R-HSA-9612973	Protein clearance, ALS/PD
Mitophagy	R-HSA-5205647	Parkin/PINK1, Parkinson’s
Tau phosphorylation	R-HSA-9646395	Alzheimer’s tauopathy
Unfolded protein response	R-HSA-381119	ER stress in neurodegeneration
NF-κB signaling	R-HSA-9006936	Neuroinflammation
Complement activation	R-HSA-166658	Synaptic pruning, early AD
mTOR signaling	R-HSA-165159	Autophagy regulation
Axon guidance	R-HSA-422475	Neurotrophic factor signaling

Relevance to SciDEX

The reactome-pathways skill in SciDEX’s Forge layer integrates with the Reactome API. Pathway enrichment results from Reactome analyses feed directly into the knowledge graph, creating links between hypothesis-relevant gene sets and pathway entities.

Cross-reference: [[mechanisms-convergent-pathways-neurodegeneration]] describes several convergent pathway themes that Reactome analysis has confirmed across neurodegeneration subtypes.

Limitations

• Human-centric: non-human model organism coverage is more limited and often inferred

• Manual curation introduces time lag; cutting-edge mechanisms may not yet be curated

• Pathway boundaries are somewhat arbitrary; gene multi-pathway membership complicates interpretation

• ORA/GSEA statistics assume gene independence, which fails for co-regulated gene clusters

• No dynamic pathway modeling; static snapshot of reaction diagrams

Cross-References

• [[mechanisms-convergent-pathways-neurodegeneration]]

• [[mechanisms-neuronal-network-dysfunction-pathway]]

• [[ai-tool-string-network]] — Protein interaction network for complementary analysis