DDA-BERT is a computational platform designed for data-dependent acquisition (DDA) proteomics analysis. It features a simple, installation-free graphical user interface (GUI) and can also run in command-line mode. Compatible with Windows and Linux environments, it is powered by the ultra-fast Sage engine—a specialized proteomics search engine for peptide identification. DDA-BERT rescores peptide identifications using a pre-trained Transformer-based model trained on 3,706 DDA files and over 95 million peptide-spectrum matches (PSMs). It demonstrates superior performance across various sample types and mass spectrometry platforms—from trace samples to mouse proteomics. Results are output in CSV format as a comprehensive summary table that is easy to manipulate and interpret, facilitating further biological insights and downstream applications.
This software is currently under development, and we welcome you to try it out. If you have any feedback or suggestions, please let us know.
Email address: guotiannan@westlake.edu.cn.
Core Features
AI driven
Adopting Utilizing advanced transformer models for precise proteomic analysis
High quality data
3706 DDA files containing over 80 million PSMs.
Friendly interface
No installation required; ready-to-use graphical user interface.
Technical specifications
Training sample size
Over 80 million PSMs
Model Architecture
Transformer-based end-to-end deep learning model
Format output
CSV table
Application scenarios
Diverse sample types, multiple mass spectrometer platforms, trace sample proteomics, and multiple species proteome data
Start using DDA-BERT
Download and experience the new generation of intelligent proteomics analysis tools now
Downloading the latest version