FoldDock

A collection of 219 heterodimers from dockground benchmark 4 dataset, and 1503 heterodimeric structures from a recent study (Green, A. G. et al. Nat. Commun. 12, 1–12 (2021)) (dubbed “marks”). The benchmark dataset was used to train an original model, that model was used to predict the structures in the second dataset. This dataset is split into three subsets, each with two splits corresponding to the groups used in the source (marks, dockground).

Quickstart Usage

Install HuggingFace Datasets package

Each subset can be loaded into python using the HuggingFace datasets library. First, from the command line install the datasets library

$ pip install datasets

Optionally set the cache directory, e.g.

$ HF_HOME=${HOME}/.cache/huggingface/
$ export HF_HOME

then, from within python load the datasets library

import datasets

Load model datasets

To load one of the RosettaMegathon2026-group3/FoldDock model datasets, use datasets.load_dataset(...):

dataset = datasets.load_dataset(f"RosettaCommons/FoldDock", "complexes")

and the dataset is loaded as a datasets.arrow_dataset.Dataset

>>> print(dataset)
DatasetDict({
    marks_complexes: Dataset({
        features: ['Uniprot ID 1', 'Uniprot ID 2', 'Sequence Length 1', 'Sequence Length 2', 'PDB', 'Chain 1', 'Chain 2', 'N residue contacts (5Ang)', 'PFAM domains 1', 'PFAM domains 2', 'Chain start 1', 'Chain end 1', 'Chain start 2', 'Chain end 2', 'Protein name 1', 'Gene name 1', 'Protein name 2', 'Gene name 2', 'Organism 1', 'Organism 2', 'Resolution (Ang)', 'contains DNA', 'contains RNA', 'Org1', 'Org2', 'kingdom1', 'kingdom2', 'Sequence 1', 'Sequence 2', 'Concatenated_Sequence'],
        num_rows: 1675
    })
    dockground_complexes: Dataset({
        features: ['Uniprot ID 1', 'Uniprot ID 2', 'Sequence Length 1', 'Sequence Length 2', 'PDB', 'Chain 1', 'Chain 2', 'N residue contacts (5Ang)', 'PFAM domains 1', 'PFAM domains 2', 'Chain start 1', 'Chain end 1', 'Chain start 2', 'Chain end 2', 'Protein name 1', 'Gene name 1', 'Protein name 2', 'Gene name 2', 'Organism 1', 'Organism 2', 'Resolution (Ang)', 'contains DNA', 'contains RNA', 'Org1', 'Org2', 'kingdom1', 'kingdom2', 'Sequence 1', 'Sequence 2', 'Concatenated_Sequence'],
        num_rows: 219
    })
})

which is a column oriented format that can be accessed directly, converted in to a pandas.DataFrame, or parquet format, e.g.

dataset.data.column('<COLUMN NAME IN DATASET>')
dataset.to_pandas()
dataset.to_parquet("dataset.parquet")

Dataset Details

Subset 1: complexes

This subset contains structural information for each of the complexes (ex. The full sequence, length of sequence, origin kingdom, etc.). Marks and dockground are combined into a single subset, but the marks set contains more comprehensive information that includes Uniprot IDs and interaction site numbers.

Subset 2: pdbs

Included in this dataset are all of the complex PDBs used in the origin work, in the pdbs directory. This subset has all of the relative paths for each of the files.

Subset 3: quality

This subset contains the quality metrics for each of the predicted protein-protein pairs.

Dataset Sources

• Repository: https://doi.org/10.17044/scilifelab.16866202.v1
• Paper: Bryant, P., Pozzati, G., & Elofsson, A. (2021). Data and most relevant results for the FoldDock project [Data set]. doi:10.17044/SCILIFELAB.16866202.V1

Citation

\@misc{https://doi.org/10.17044/scilifelab.16866202.v1,
  doi = {10.17044/SCILIFELAB.16866202.V1},
  url = {https://figshare.scilifelab.se/articles/dataset/Data_and_most_relevant_results_for_the_FoldDock_project/16866202/1},
  author = {Bryant,  Patrick and Pozzati,  Gabriele and Elofsson,  Arne},
  keywords = {Bioinformatics and computational biology not elsewhere classified},
  title = {Data and most relevant results for the FoldDock project},
  publisher = {Stockholm University},
  year = {2021},
  copyright = {Apache 2.0}
}

Dataset Card Authors

• Katie Sie katiesie/@/uw.edu
• John Bickel johndjbickel/@/gmail.com
• Ryan Mauder rmauder/@/butler.edu
• Corleigh Forrester corleighforr/@/gmail.com

Acknowledgements: Patrick Bryant, Gabriele Pozzati, Arne Elofsson
License: apache-2.0