🧬 DeepRank3

A Deep Learning Method for Ranking Protein Structural Models

---

🚀 Overview

DeepRank3 is a deep learning-based system designed for ranking and quality assessment of protein structural models.
It integrates multiple deep learning modules, tools, and databases to estimate the accuracy of protein tertiary and quaternary structures.

---

📦 1. Download and Install DeepRank3

⚠️ Use a short installation path to avoid file path length issues.

git clone https://github.com/jianlin-cheng/DeepRank3.git
# If cloning fails, try with your GitHub username:
git clone https://huge200890@github.com/jianlin-cheng/DeepRank3.git

cd DeepRank3

wget "https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-$(uname)-$(uname -m).sh"
bash Miniforge3-$(uname)-$(uname -m).sh  # accept all terms and install to the default location
rm Miniforge3-$(uname)-$(uname -m).sh  # (optionally) remove installer after using it
source ~/.bashrc  # alternatively, one can restart their shell session to achieve the same result

---

🧰 2. Setup Database and Tools

a. Install python2.7 environment

mamba create -n python27 python=2.7
mamba activate python27

b. Edit setup_database.pl

1. Create a directory for databases:

   mkdir -p /data/commons/DeepRank_db_tools/

2. Set the variable $DeepRank_db_tools_dir in setup_database.pl:

   $DeepRank_db_tools_dir = "/data/commons/DeepRank_db_tools/";

c. Run setup

perl setup_database.pl

📘 Please refer to cite_methods_for_publication.txt for citation guidelines.
All external tools can also be downloaded from their respective official websites.

---

⚙️ 3. Configure DeepRank3

a. Edit configure.pl

Set the same database path used above:

$DeepRank_db_tools_dir = "/data/commons/DeepRank_db_tools/";

b. Save and run:

perl configure.pl

---

🐍 4. Verify Python Environment

Activate the virtual environment:

source DeepRank_db_tools/tools/python_virtualenv/bin/activate

If it fails, manually install the environment:

sh installation/DeepRank_manually_install_files/P4_python_virtual.sh

---

🔧 5. Configure Keras Backend

Set Theano as the backend:

mkdir -p ~/.keras
vi ~/.keras/keras.json

Paste:

{
    "epsilon": 1e-07,
    "floatx": "float32",
    "image_data_format": "channels_last",
    "backend": "theano"
}

---

🧠 6. Install DeepDist Tool (Python 3.6)

#Install Python 3.6 environment:
mamba create -n python36 python=3.6
mamba activate python36

cd tools/deepdist
python setup.py
python configure.py
sh installation/set_env.sh

---

📊 7. Install DistRank Tool

mamba activate python36
cd ../DistRank
mkdir env
python configure.py
sh installation/set_env.sh

---

🧩 8. Run DeepRank3 for Quality Assessment

🔹 Usage

sh bin/DeepRank3_Cluster.sh <target_id> <fasta_file> <model_dir> <output_dir>

🔹 Examples

mamba activate python27

# Cluster-based ranking
sh bin/DeepRank3_Cluster.sh T0953s1 examples/T0953s1.fasta examples/T0953s1 examples/test_out

# Single-model QA
sh bin/DeepRank3_SingleQA.sh T0953s1 examples/T0953s1.fasta examples/T0953s1 examples/test_out

# Lightweight QA
sh bin/DeepRank3_SingleQA_lite.sh T0953s1 examples/T0953s1.fasta examples/T0953s1 examples/test_out

---

🧾 Citation

If you use DeepRank3 or its components in your research, please cite the corresponding methods listed in:

@article{liu2022improving,
  title={Improving protein tertiary structure prediction by deep learning and distance prediction in CASP14},
  author={Liu, Jian and Wu, Tianqi and Guo, Zhiye and Hou, Jie and Cheng, Jianlin},
  journal={Proteins: Structure, Function, and Bioinformatics},
  volume={90},
  number={1},
  pages={58--72},
  year={2022},
  publisher={Wiley Online Library}
}

cite_methods_for_publication.txt

---

📜 License

This project is released under the MIT License.