EnGens: a computational framework for generation and analysis of representative protein conformational ensembles

Repository for the computational framework for generation and analysis of representative protein conformational ensembles.

Citation

@article{10.1093/bib/bbad242,
    author = {Conev, Anja and Rigo, Mauricio Menegatti and Devaurs, Didier and Fonseca, André Faustino and Kalavadwala, Hussain and de Freitas, Martiela Vaz and Clementi, Cecilia and Zanatta, Geancarlo and Antunes, Dinler Amaral and Kavraki, Lydia E},
    title = "{EnGens: a computational framework for generation and analysis of representative protein conformational ensembles}",
    journal = {Briefings in Bioinformatics},
    volume = {24},
    number = {4},
    pages = {bbad242},
    year = {2023},
    doi = {10.1093/bib/bbad242}
}

---

Demo

Try running our demos on Google Colab:

• Dynamic Use-Case (with a custom MD trajectory as input)
  • 
  • This demo features an MD test example of a small pentapeptide
  • You can also upload your own trajectory!
• Static Use-Case (with UniProt ID as input)
  • 
  • This demo features the P03372 UniProt accession related to the ER-alpha estrogent receptor
  • But you can try it out with your own choice of UniProt ID too!

Try runnning our notebooks on Binder:

Documentation

Check out the detailed readthedocs documentation here: https://engens.readthedocs.io/en/latest/

Installation instructions

Docker image pull

The prefered and easiest is by pulling the docker image made available publicly.

prerequisites: docker

Just pull the image:

docker pull ac121/engens:latest

You're all set!

Note: this step should not take longer than 15min. On Windows the PowerShell sometimes gets stuck - do a right click in the terminal to check the progress after 10-15min.

For other installation options check the section Advanced Installation bellow.

Running EnGens

Linux: run the following command from the working directory with this code

docker run -it --rm -v $(pwd):/home/engen/ -p 8888:8888 ac121/engens:latest jupyter notebook --ip=0.0.0.0 --port=8888

Windows: run the following command from the working directory with this code

docker run -it --rm -v ${pwd}:/home/engen/ -p 8888:8888 ac121/engens:latest jupyter notebook --ip=0.0.0.0 --port=8888

Using these commands a link will pop up (something like http://127.0.0.1:8888/?token=7f4fb1ded621bda931880bd3cd1c62431d47abfbb91116ac). Follow this link and you will find a set of notebooks:

• Workflow1-crystal_structures.ipynb
• Workflow1-FeatureExtraction.ipynb
• Workflow2-DimensionalityReduction.ipynb
• Workflow3-Clustering.ipynb
• Workflow4-ResultSummary.ipynb

For static workflow start with Workflow1-crystal_structures.ipynb and continue with Workflow2-4. For dynamic workflow start with Workflow1-FeatureExtraction.ipynb and continue with Workflow2-4.

---

Code

All the code and classes used in the notebooks are found in the directory ./EnGeNs/engens_code/engens/core/

Advanced Installation

1. Docker image build

You can clone this repo and build the docker image yourself.

prerequisites: git, docker

1. Clone the github repo:

git clone https://github.com/KavrakiLab/EnGens.git

2. Build the image:

cd EnGens
docker build -t test_engens:latest .

You're all set!

2. Conda environment build

If you don't want to use docker, you can clone this repo and install using conda (or mamba which will be faster).

prerequisites: conda or mamba

1. Clone the github repo:

git clone https://github.com/KavrakiLab/EnGens.git

2. Install with conda (or mamba)

cd EnGens
conda env create -f ./environment.yaml
#mamba env create -f ./environment.yml

conda activate engens
#mamba activate engens

./linux_setup.sh
#or ./windows_setup.sh

If the command ./linux_setup.sh fails due to not having pypatch - do pip install pypatch.

Acknowledgements

EnGens relies on and/or references the following separate libraries and packages:

• Structural bioinformatics (and MD) software
  • MDTraj
  • MDAnalysis
  • PyTraj
  • MSMTools
  • PyEmma
  • OpenMM
  • HDE
  • mTM-align
  • BioPython
• Visualization
  • NGLViewer
  • py3Dmol
  • ProtVista
• General ML tools
  • scikit-learn
  • scipy
  • umap-learn
• Others
  • plotly
  • pandas
  • numpy

We thank all their contributors and maintainers!

Funding

Work on this project by A.C. and L.E.K. has been supported in part by the National Institutes of Health NIH [U01CA258512]. Other support included: University of Edinburgh and Medical Research Council [MC_UU_00009/2 to D.D.]; Computational Cancer Biology Training Program fellowship [RP170593 to M.M.R.]; The Brazilian National Council for Scientific and Technological Development [CNPq no. 440412/2022-6 to G.Z.]; University of Houston Funds and Rice University Funds.

---