iPTMnet

Description

iPTMnet is an integrated resource for protein post-translational modification (PTM) network discovery that employs an integrative bioinformatics approach combining text mining, data mining, and ontological representation to capture rich PTM information, including PTM enzyme-substrate-site relationships, PTM-specific protein-protein interactions (PPIs), and PTM conservation across species.

Coverage

As of Release 4.1 (August 2017), iPTMnet contains:

• 654,500+ unique PTM sites in over 62,100 modified proteins
• 1,200 PTM enzymes
• 12,700 distinct enzyme-substrate pairs
• 24,300 distinct enzyme-substrate-site combinations
• 1,470 PTM-dependent protein-protein interactions
• 30,500 publications describing PTM and/or PPI relations

PTM Types Covered

Eight major PTM types:

1. Phosphorylation (primary)
2. Ubiquitination
3. Acetylation
4. Methylation
5. Glycosylation
6. S-nitrosylation
7. Sumoylation
8. Myristoylation

Organisms

Top organisms represented:

• Human
• Mouse
• Rat
• Arabidopsis thaliana
• Saccharomyces cerevisiae
• Schizosaccharomyces pombe

Data Sources

Text Mining Systems

• RLIMS-P: Rule-based information extraction system for literature mining of protein phosphorylation information from PubMed abstracts and full-length articles
• eFIP: Full-scale mining of PubMed Central Open Access articles for phosphorylation information

Curated Databases (11 sources integrated)

1. PhosphoSitePlus (PSP): Phosphorylation, ubiquitination, acetylation, methylation (human, rat, mouse)
2. Phospho.ELM: Phosphorylation sites in animal proteins
3. PhosPhAt: Mass spectrometry phosphorylation sites in Arabidopsis thaliana
4. PhosphoGrid: In vivo phosphorylation sites in Saccharomyces cerevisiae
5. PomBase: Fission yeast (Schizosaccharomyces pombe) comprehensive database
6. UniProtKB: Comprehensive protein database with expert-annotated PTM features
7. P3DB: Plant protein phosphorylation data
8. neXtProt: Human protein knowledgebase with kinase focus
9. HPRD: Human protein PTMs and enzyme-substrate relationships
10. Signor: Causal relationships between biological entities including PTM-enzyme substrate relations
11. dbSNO: Experimentally verified cysteine S-nitrosylation sites

Ontological Framework

• Protein Ontology (PRO): Organizes proteins and PTM proteoforms with hierarchical representation (family→gene→sequence→modification) enabling representation of experimentally validated PTM combinations

Key Features

Unique Capabilities

1. Proteoform Representation: Shows experimentally validated combinations of PTMs on proteins (unique feature)
2. PTM Conservation Analysis: Alignment of orthologous proteoforms across species enabling PTM site and proteoform prediction
3. Confidence Scoring: Quality scores (0-4 stars) for PTM information based on source quality, multiple source support, and publication evidence
4. Network Visualization: Cytoscape-based PTM sites and proteoforms as nodes with enzyme-site and PPI relationships as edges
5. Integrated Text Mining: Up-to-date PTM information from automated monthly processing of all PubMed abstracts and PMC Open Access full-text articles

Search & Browse Functionality

• Search by UniProtKB AC/ID, protein/gene name, PRO ID, PMID
• Restrict by PTM type, enzyme/substrate role, organism
• Batch Retrieval: Process up to 500 PTM sites at once to obtain PTM enzyme or PTM-dependent PPI information
• Literature Search: Dual search modes (protein database search + phosphorylation literature search)
• Interactive Entry Reports with multiple data tables (substrate, PTM enzyme, proteoform, PPI)

Visualization Tools

1. Cytoscape Network View: Interactive network visualization of PTM enzyme-substrate-site, proteoform-site, and PPI relationships
2. Sequence Alignment Viewer: Multiple sequence alignment (MUSCLE algorithm) showing PTM conservation across species with color-coded modifications
3. Overlapping PTMs: Highlights residues with multiple modification types (potential PTM cross-talk sites)

RESTful API

• Programmatic access to iPTMnet data
• Documented in Methods Mol Biol (2022) and Database (Oxford) (2020)
• Enables automated integration with external analysis workflows

Quality Control

• Monthly updates of text mining results
• Integrity checks on kinase information (verification against UniProtKB ‘kinase’ keyword)
• PTM type validation (conformance to known residue types)
• PTM site sequence validation (residue at expected position in UniProtKB sequence)
• Monitoring for retracted articles with correction/removal of affected data

Applications

• Functional interpretation of phosphoproteomic data
• Kinase signaling pathway analysis (e.g., connecting phosphosites from mass spec experiments to kinase pathways)
• PTM-dependent PPI discovery
• Cross-species PTM conservation studies
• Drug target identification (e.g., EGFR inhibitor erlotinib response analysis)
• Hypothesis generation for PTM cross-talk and proteoform biology

Maintenance & Support

• Maintained by Protein Information Resource (PIR)
• University of Delaware: 15 Innovation Way, Suite 205, Newark, DE 19711
• Georgetown University Medical Center: 3300 Whitehaven Street, NW, Suite 1200, Washington, DC 20007
• Funding: NSF grant ABI-1062520, NIH/NIGMS grant U01GM120953
• Website traffic: >6 million hits from >16,000 unique IPs (first half of 2017)

Technical Implementation

• Database: Oracle 12c release 1, dimensional model design
• Front-end: Django (Python Web Framework)
• Visualization: Cytoscape.js (version 2.4.2 graph theory library)
• Sequence Alignment: MUSCLE algorithm
• Text Mining: PubTator for NCBI gene ID mapping
• ID Mapping: UniProt Protein ID and gene name mapping tools

Access & Documentation

• Main Portal: https://research.bioinformatics.udel.edu/iptmnet/
• Statistics: http://research.bioinformatics.udel.edu/iptmnet/stat
• Help Documentation: Available online (PDF)
• Tutorial: http://research.bioinformatics.udel.edu/iptmnet/static/iptmnet/files/iPTMnet_Help.pdf
• API Documentation: See publications PMID:35696082 and PMID:32395768
• Europe PubMed Central Integration: External link tab connection for literature cross-referencing