Software for data-independent acquisition (DIA) data analysis with deep representation features and FDR-controlled match-between runs (DreamDIAlignR).

New Feature of DreamDIA3: DreamDIAlignR

DreamDIAlignR is a novel cross-run peptide-centric analysis workflow that allows for consistent cross-run peak picking and FDR-controlled peak scoring.

1. Docker image & quick example

We recommend using Docker containers to simplify installation and avoid complex compilation processes.

1. Pull the docker image from DockerHub.

docker pull mingxuangao/dreamdia:v3.2.0

2. Start a container.

docker run -it --name dreamdia_example --gpus all -v /YOUR/OWN/WORK/PATH:/tmp/work mingxuangao/dreamdia:v3.2.0 /bin/bash

• The --gpus all argument enables GPU support within the container, which is essential for running the latest version of DreamDIA.

3. Activate the conda environment for DreamDIA.

conda activate dreamdia

4. (optional) Test the availability of GPUs.

python /root/check_gpus.py

5. Run the demo.

# Enter the working directory
cd

# Run DreamDIA peptide peak identification module
python DreamDIA/DreamDIA.py dreamscore --file_dir example_data/raw_data --lib example_data/Spyogenes_library.tsv --out dreamscore_out

# Run DreamDIAlignR
python DreamDIA/DreamDIA.py dreamprophet --dream_dir dreamscore_out --out dreamdia_out --dreamdialignr --r_home /root/miniconda3/envs/dreamdia/bin/R 

2. Build from source

Requirements

Linux
python >= 3.6.0
pyteomics
numpy
pandas
seaborn
cython
scikit-learn
tensorflow
keras-gpu >= 2.4.3
statsmodels
xgboost
networkx
rpy2 >= 3.5
R >= 4.2.0

If .raw files are to be directly input into DreamDIA on Linux systems, mono must be installed.

We recommend using Anaconda to set up the environment and install the necessary libraries as outlined below.

# Initiate a conda virtual environment called "dreamdia"
conda create -n dreamdia python=3.6.12

# Activate the "dreamdia" virtual environment
conda activate dreamdia

# Install the libraries
conda install -y keras-gpu
conda install -y scikit-learn
conda install -y py-xgboost-cpu
conda install -y cython
conda install -y seaborn
conda install -y statsmodels
conda install -y pyteomics -c bioconda
conda install -y networkx
conda install -y r-base=4.2.1 -c conda-forge
conda install -y rpy2

Download

https://github.com/xmuyulab/DreamDIA/releases/tag/v3.2.0

Installation

cd DreamDIA
bash build.sh

3. Quick start

The latest version of DreamDIA analysis workflow consists of two steps: dreamscore and dreamprophet.

# Step1
python DreamDIA-vXXX/DreamDIA.py dreamscore --help

# Step2
python DreamDIA-vXXX/DreamDIA.py dreamprophet --help

dreamscore identifies peptide peaks for each provided run ,while dreamprophet needs the output of the first step and performs optional match-between-runs and statistical analysis.

dreamprophet does not perform match-between-runs by default. However, if the --dreamdialignr and --r_home options are specified, the DreamDIAlignR algorithm will be activated and executed.

4. Notes

1. DIA raw data files

• Centroided .mzML or .mzXML files are supported at any time.
• If .raw files are going to be fed directly to DreamDIA on Linux systems, mono must be installed first for the data format conversion by ThermoRawFileParser.

All raw data files should be in one folder as shown below.

# rawdata_dir/
	rawdata_1.mzML
	rawdata_2.mzXML
	rawdata_3.raw

2. Spectral libraries

Only .tsv libraries are supported. All of the columns required by DreamDIA are listed in DreamDIA/lib_col_settings. Users can modify this file to adjust their own spectral libraries.

3. output

DreamDIA outputs peptide and protein identification and quantification results. An empty directory is suggested for the --out argument to save all of the output files.

4*. Advanced: train your own deep representation models

See the guidance in Train_customized_models.ipynb to train your own deep representation models.

5. Cite this article

[1] Gao, M., Yang, W., Li, C. et al., Deep representation features from DreamDIA^XMBD improve the analysis of data-independent acquisition proteomics. Commun Biol 4, 1190 (2021). https://doi.org/10.1038/s42003-021-02726-6

[2] Gao, M., Gupta, S., Yang, W. et al., Scoring information integration with statistical quality control enhanced cross-run analysis of data-independent acquisition proteomics data. BioRxiv, 2024. https://www.biorxiv.org/content/10.1101/2024.12.19.629475v1