dynamicPDB

Dynamic PDB: A New Dataset and a SE(3) Model Extension by Integrating Dynamic Behaviors and Physical Properties in Protein Structures

Overview

Dynamic PDB is a large-scale dataset that enhances existing prestigious static 3D protein structural databases, such as the Protein Data Bank (PDB), by integrating dynamic data and additional physical properties. It contains approximately 12.6k filtered proteins, each subjected to all-atom molecular dynamics (MD) simulations to capture conformational changes.

Compared with previously existing protein MD datasets, dynamic PDB provides three key advancements:

1. Extended simulation durations: Up to 1 microsecond per protein, facilitating a more comprehensive understanding of significant conformational changes.
2. Finer-grained sampling intervals: 1 picosecond intervals, allowing for the capture of more detailed allosteric pathways.
3. Enriched array of physical properties: Captured during the MD process, including atomic velocities and forces, potential/kinetic energies, and the temperature of the simulation environment, etc.

What dynamic PDB contains?

The attributes contained in dynamic PDB are listed as follows:

In addition, the following data are stored during the MD simulation:

Download Dataset

You can easily get dynamic PDB dataset from our ModelScope repo (100 ns or the complete version). These datasets are stored in compressed file format, and their usage is shown below. We also provide a ready-to-use lightweight version, which is a downsampled variant of the 100 ns dataset with a 1 ns frame interval.

1. Make sure you have Git LFS installed:

sudo apt-get install git-lfs
# Initialize Git LFS
git lfs install

1. Navigate to your DATA_ROOT and clone the source:

GIT_LFS_SKIP_SMUDGE=1 git clone https://www.modelscope.cn/datasets/fudan-generative-vision/dynamicPDB.git dynamicPDB_raw

GIT_LFS_SKIP_SMUDGE=1 configures Git to clone the pointers for all LFS files.

1. Download data with a specific protein_id, for example 1a62_A:

cd dynamicPDB_raw
git lfs pull --include="{protein_id}/*"

1. Merge the split-volume compression into one file and then unzip the .tar.gz file:

cat {protein_id}/{protein_id}.tar.gz.part* > {protein_id}/{protein_id}.tar.gz
cd ${Your Storage Root}
mkdir dynamicPDB  # ignore if directory exists
tar -xvzf dynamicPDB_raw/{protein_id}/{protein_id}.tar.gz -C dynamicPDB

Finally, the dataset should be organized as follows:

…

For ease of use, we have also provided segmented versions of the data (directories {protein_id}_dcd, {protein_id}_T, {protein_id}_F, and {protein_id}_V), each representing one-tenth of the total simulation duration, sequentially named from 0 to 9 in chronological order. The files {protein_id}_T.dcd, {protein_id}_T.pkl, {protein_id}_F.pkl, {protein_id}_V.pkl are their corresponding combination.

Applications

1. Physics condition: we extend the SE(3) diffusion model to incorporate sequence features and physical properties for the task of trajectory prediction.
2. 4D diffusion: we introduce a 4D diffusion with motion alignment to learn dynamic protein structures.
3. Conformation sampling: we introduce a 4D diffusion to sample diverse protein conformations.

Acknowledgements

We would like to thank the contributors to the OpenFold, OmegaFold.
If we missed any open-source projects or related articles, we would like to complement the acknowledgement of this specific work immediately.