PTMVision

PTMVision is an interactive web-based platform for the visual exploration of post-translational modifications (PTMs) identified in mass spectrometry-based proteomics data. It enables researchers to interpret the complex PTM landscape, supporting insights into molecular mechanisms of biological processes and disease.

🔗 Web server: https://ptmvision-tuevis.cs.uni-tuebingen.de

✨ Features

• Rich interactive visualizations for sample-level and protein-level PTM data
• Linked plots for seamless navigation and exploration
• 3D structure integration and contact map analysis for structural context
• Highlights residue-residue PTM interactions in close spatial proximity
• Annotation integration from UniProt and UniMod for interpretability
• Support for multiple search engine output formats and a custom CSV format
• Session export/import for reproducibility and collaboration
• Intuitive UI accessible to users without programming expertise

📁 Supported Input Formats

PTMVision accepts output from several widely-used search engines:

| Search Engine | Tested Versions | Key Columns | Processing Steps | | ------------------------------------------ | ---------------------------------- | ------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------- | | ionbot | v0.10.0, v0.11.0 | uniprot_id, unexpected_modification, position | Direct parsing of modification information | | MSFragger (via FragPipe + PTMShepherd) | MSFragger v4.0, PTMShepherd v2.0.6 | Modified Peptide, Assigned Modifications, Localization, Protein ID, Protein Start | Filtering ambiguous or multi-localized sites; mass shift mapping to UniMod; UniProt sequence alignment; UniMod classification | | Sage | v0.13.1 | See psm-utils | FDR filtering; decoy removal; UniMod mapping; PTM localization from variable mods only | | MaxQuant | v2.4.13.0 | Search result columns used for mapping | Mapping MaxQuant-specific names to UniMod; resolving multi-mapping peptides; UniProt mapping | | Spectronaut (PTM Site Report) | v9 | PTM.ProteinId, PTM.SiteLocation, PTM.ModificationTitle, PTM.SiteAA | Infers mass shifts and UniMod classifications from reported data and UniProt sequence |

Note: For FragPipe-derived data, localization confidence is inferred from the MSFragger Localization field. Use caution until future versions provide dedicated localization scores.

📄 Custom Plain CSV Format

Users can submit manually compiled PTM data using a structured CSV format when other formats are unavailable.

Required Columns

| Column Name | Description | | ----------------------------- | ------------------------------------------------------------------------ | | uniprot_id | UniProt accession of the protein | | position | 1-based index of the modified residue | | modification_unimod_name | Name of the UniMod modification | | modification_unimod_id | UniMod ID | | mass_shift (optional) | Monoisotopic mass shift in Daltons (inferred from UniMod ID if omitted) | | classification (optional) | UniMod classification (inferred from ID and sequence context if omitted) |

Example entry:

P04075,5,carbamidomethyl,4,57.021464,Artefact

⚠ Note: Classification relies on up-to-date UniProt sequences. Sequence version mismatches may lead to misaligned positions.

📄 Download CSV template

🐳 Running Locally via Docker

To deploy PTMVision locally:

```bash

Clone repository

git clone https://github.com/Integrative-Transcriptomics/PTMVision.git cd PTMVision

Build Docker image

docker build --tag ptmvision app

Run locally

docker run -dp 127.0.0.1:5001:5001 ptmvision ```

Note: Prebuilt Docker images are not currently provided.

🙋 Need Help?

• 📖 Read the Usage Guide
• 🐛 Found an issue or missing format? Open a GitHub issue

✏️ Citation

If you use this software, please cite it as below:

``` PTMVision: An Interactive Visualization Webserver for Post-translational Modifications of Proteins

Simon Hackl, Caroline Jachmann, Mathias Witte Paz, Theresa Anisja Harbig, Lennart Martens, and Kay Nieselt

Journal of Proteome Research 2025 24 (2), 919-928

DOI: 10.1021/acs.jproteome.4c00679 ```