DDA-BERT is an end-to-end deep learning tool for rescoring peptide-spectrum matches (PSMs) in data-dependent acquisition (DDA) proteomics. Built on a Transformer-based architecture and trained on 12,285 DDA-MS files comprising approximately 271 million high-confidence PSMs, DDA-BERT effectively models the complex relationships between peptide sequences and tandem mass spectra. It supports rescoring results generated from a single database search engine as well as integrated outputs from multiple search engines.
DDA-BERT demonstrates robust and consistent performance across diverse biological systems, including animal, plant, and microbial proteomes. When applied to HLA immunopeptidomics datasets, DDA-BERT continues to show strong and reliable performance. It is applicable to low-input contexts such as trace-level and single-cell proteomics, offering a scalable and reliable solution for improving peptide identification in mass spectrometry-based workflows.
Detailed installation and usage instructions are available in the README under the software/ directory (https://github.com/zhiyuajun/DDA-BERT/tree/main/software). The portable executable can be downloaded from https://guomics.com/software/DDA-BERT.
Hardware Requirements:
• Operating System: Compatible with Linux-based operating systems.
• Processor: A dual-core processor is recommended; the platform can also run on a single-core processor.
• Memory: At least 40 GB of RAM is recommended. Higher memory configurations are advised for processing large-scale mass spectrometry or FASTA datasets.
• Graphics Processing Unit (GPU): An NVIDIA GPU that supports bfloat16 (bf16) precision inference is required. CUDA support is necessary, and a minimum of 20GB GPU memory is recommended.
This repository is organized into the following main directories, each serving a distinct role in the DDA-BERT workflow:
-
software/
This directory contains the core implementation of the DDA-BERT identification and rescoring pipeline, including all required source code, executable scripts, and configuration files necessary to run the software.
-
training_eval/
The
training_evaldirectory contains the code used for model training and evaluation. It includes scripts for configuring training parameters, executing training jobs, and performing systematic evaluation of trained models. -
scripts/
This directory contains auxiliary scripts for data processing, result summarization, and visualization. These utilities support downstream analysis, such as plotting performance metrics and generating figures for method evaluation and result interpretation.
-
demo_data/
For demonstration and testing purposes, the
demo_datadirectory includes example datasets that allow users to quickly validate the software functionality and reproduce the complete analysis workflow without preparing custom input data. Additional demo datasets in other formats are available for download from Google Drive (https://drive.google.com/drive/folders/1pDxTuFYKoy-uJmq1QS7mWBuz6pkVh3jN).
Evaluation typically completes in ~20 minutes, depending on the number of spectra and available GPU/CPU resources.
For the provided test data, the complete workflow, including database search, PSM rescoring, and protein inference, takes approximately 17.5 minutes on a single NVIDIA A100 (40GB) GPU with 20 CPU cores (AMD EPYC 7742 64-Core Processor).
Currently, the tool supports input data in .mzML and Bruker .d formats. Sciex .wiff files are not directly supported and must be converted to .mzML prior to analysis. The source code is openly available and modifiable (see the license for details), allowing users to extend the tool to accommodate additional data formats if needed.
Benchmarking results comparing DDA-BERT with other tools, along with the corresponding FASTA files, trained models, and database search configuration files, are available at: https://drive.google.com/drive/folders/1pDxTuFYKoy-uJmq1QS7mWBuz6pkVh3jN
Results are stored in CSV format. Detailed descriptions and interpretation of the output files are provided in the user manual at https://guomics.com/software/DDA-BERT/downloads.html.
For questions regarding usage or licensing, please contact Dr. Guo via email: guotiannan@westlake.edu.cn www.guomics.com