Eukaryotic Phosphorylation Site Database 2.0

Overview

The Eukaryotic Phosphorylation Site Database (EPSD) is a comprehensive resource for protein phosphorylation data across eukaryotic species, providing experimentally validated phosphorylation sites with detailed functional annotations. Originally released in 2019 as EPSD 1.0, the database has undergone significant expansion in version 2.0 (published July 2025) to become one of the largest and most extensively annotated phosphorylation databases available.

Data Content and Coverage

EPSD 2.0 contains 2,769,163 experimentally identified phosphorylation sites occurring on 362,707 phosphoproteins from 223 eukaryotic species, representing a 71% increase in phosphorylation sites compared to version 1.0. The species coverage includes:

• 95 animals (including human, mouse, rat, Drosophila, C. elegans)
• 20 protists
• 61 plants (including Arabidopsis thaliana)
• 48 fungi (including S. cerevisiae, S. pombe)

Among all phosphorylation sites, the database contains:

• 1,930,151 (69.70%) phosphoserine (pS) residues
• 638,944 (23.07%) phosphothreonine (pT) residues
• 200,020 (7.22%) phosphotyrosine (pY) residues
• Additional sites on other residues including histidine

The database encompasses 3,364,760 phosphopeptides with their primary references, preserving the initial peptides detected via mass spectrometry along with cell/tissue origins when available.

Functional Annotations

A distinguishing feature of EPSD 2.0 is its comprehensive functional annotation:

• 88,074 functional downstream events annotated for 32,762 phosphorylation sites
• 58 categories related to protein function effects (e.g., enzymatic activity, molecular association, subcellular localization)
• 107 categories about biological process impacts (e.g., cell growth, disease progression, autophagy)

For eight model organisms (human, mouse, rat, Drosophila melanogaster, Caenorhabditis elegans, Arabidopsis thaliana, Schizosaccharomyces pombe, and Saccharomyces cerevisiae), EPSD 2.0 provides multilayer annotations integrated from 100 external resources covering:

• Kinase/phosphatase regulators
• Transcription regulators
• Three-dimensional protein structures (PDB and AlphaFold)
• Physicochemical characteristics
• Genomic variations
• Functional descriptions
• Protein domains
• Molecular interactions (protein-protein interactions)
• Drug-target associations
• Disease-related data
• Orthologs across species
• Transcript expression levels
• Proteomics data
• Subcellular localization
• Regulatory pathways

Data Collection and Quality

EPSD 2.0 data is derived from multiple sources:

• 873,718 new phosphorylation sites from 575 high-throughput phosphoproteomic studies collected from PubMed literature
• Integration with EPSD 1.0 (1,616,804 sites in 209,326 proteins from 68 species)
• 362,190 additional sites from 10 public databases (PhosphoSitePlus, dbPTM, UniProt, PhosPhAt, BioGRID, RegPhos, iPTMnet, Plant PTM Viewer, Pf-Phospho, Scop3P)

All phosphorylation sites are mapped to UniProt reference sequences with exact position information. For sites from high-throughput experiments, localization probability (LP) scores are provided and classified into four classes:

• Class I: LP > 0.75
• Class II: 0.50 < LP ≤ 0.75
• Class III: 0.25 ≤ LP ≤ 0.50
• Class IV: LP < 0.25

Bioinformatic Features

EPSD 2.0 provides several computational analyses for each phosphorylation site:

• Disorder propensity calculated by IUPred3 (phosphorylation sites are predominantly situated in intrinsically disordered regions)
• Surface accessibility calculated by NetSurfP-3.0 (residues with higher surface accessibility are more likely to be phosphorylated)
• 3D structure visualization using 3Dmol.js with phosphorylation sites highlighted on experimental structures (PDB) or predicted structures (AlphaFold)

Database Interface and Search Functions

The EPSD 2.0 web portal offers multiple search and browse options:

• Substrate Search: Search by protein name, gene name, UniProt ID, EPSD ID, or species
• Peptide Search: Search for phosphopeptides (e.g., “KKpTLCGTPNYIAPEVLSK” where “p” indicates phosphorylated residue)
• Advanced Search: Multiple condition queries for precise searches
• Batch Search: Line-by-line input of multiple search terms
• BLAST Search: Sequence similarity search to identify identical or homologous proteins
• Browse by Species: Navigate through all 223 supported eukaryotic species

Each protein entry displays:

• Basic information (UniProt ID, gene name, organism, NCBI taxa ID)
• 3D structure with highlighted phosphorylation sites
• Comprehensive list of all phosphorylation sites with disorder/accessibility predictions
• Detailed phosphopeptide information with LP scores and references
• Functional annotations and downstream effects
• Integrated annotations across 15 categories

Applications and Impact

EPSD 2.0 serves multiple research communities:

• Phosphorylation research: Comprehensive catalog of experimentally validated sites
• Kinase-substrate relationships: Identification of upstream kinases and downstream effects
• Machine learning: Training data for phosphorylation site predictors (e.g., GPS 6.0 developed using EPSD data)
• Disease research: Understanding aberrant phosphorylation in cancer, neurodegeneration, and cardiovascular disorders
• Drug discovery: Identifying phosphorylation-related therapeutic targets
• Comparative phosphoproteomics: Cross-species analysis of phosphorylation patterns

The database demonstrates that 71.87% of phosphoproteins are modified at multiple sites, highlighting multisite phosphorylation as a dominant regulatory mechanism.

Maintenance and Updates

EPSD 2.0 is actively maintained by the CUCKOO Workgroup at Huazhong University of Science and Technology, China, led by Professor Yu Xue. The database was last updated on November 1, 2025, and is regularly maintained with:

• Continuous collection of newly reported phosphorylation sites from literature
• Expansion of functional annotations through literature review
• Integration of additional external resources
• Updates to structural data and annotations

The total database size is approximately 36.2 GB, representing a 2.5-fold increase compared to EPSD 1.0 (14.1 GB).

Data Access and Download

All phosphorylation sites and annotation datasets in EPSD 2.0 are freely accessible for download at https://epsd.biocuckoo.cn/Download.php. The database has been submitted to Database Commons at the National Genomics Data Center (NGDC), China National Center for Bioinformation (CNCB), and is publicly accessible.

Technical Details

• Database identifiers: Each phosphoprotein receives an automatically generated EPSD ID (EP-) with UniProt accession as alternative identifier
• Primary references: PubMed IDs provided for all sites with experimental evidence
• Data format: Structured data with comprehensive metadata
• API access: Available through the download portal
• Updates: Regular maintenance with last update November 1, 2025

Related Resources

EPSD 2.0 integrates and complements data from multiple phosphorylation databases including PhosphoSitePlus, dbPTM, UniProt, PhosPhAt, iPTMnet, and others, providing one of the most comprehensive views of the eukaryotic phosphoproteome. The database also supports the GPS (Group-based Prediction System) 6.0 web server for kinase-specific phosphorylation site prediction.