Datasets: AFsample2 predicts multiple conformations and ensembles with AlphaFold2

Directory Overview

An overview of the directory structure, along with a description of each folder and its contents. Extract as follows.

tar --use-compress-program=unzstd -xvf analysis_results.tar.zst

tar --use-compress-program=unzstd -xvf input_datasets.tar.zst

tar --use-compress-program=unzstd -xvf generated_models.tar.zst

1. analysis_results

1. Contains results (.csv) of ensemble analysis
2. Files required to generate images for the manuscript with included summarize.ipynb notebook (updated notebook available at https://github.com/iamysk/AFsample2/blob/main/notebooks/summarize.ipynb)

tar --use-compress-program=unzstd -xvf analysis_results.tar.zst

└── analysis_results
   ├── general
   ├── oc23
   │   ├── afsample2
   │   ├── SPEACH_AF
   │   ├── ...
   └── tp16
       ├── afsample2
       ├── SPEACH_AF
       ├── ...

2. generated_models

1. This directory stores models generated during the project.
2. It is organized by dataset names (e.g., oc23, tp16).

Inside generated_models:

oc23 and tp16: Subdirectories corresponding to models corresponding to various methods (AFcluster, AFsample, AFsample2, MSA subsampling, SPEACH_AF) employed in the study. All methods except AFcluster have been implemented in AFsample2 code.

tar --use-compress-program=unzstd -xvf input_datasets.tar.zst

└── input_datasets
    ├── oc23
    │   ├── fastas
    │   ├── filtered_dict.pickle    # pdbids and stats for states
    │   ├── msas                               # in .pkl format
    │   └── pdbs                               # All reference pdbs used in the study
    └── tp16
        ├── fastas
        ├── msas
        └── pdbs

3. input_datasets

1. Contains raw input datasets used for analysis and model generation.
2. Organized by dataset names (e.g., oc23, tp16).

Inside input_datasets (oc23 and tp16):

• • fastas: Directory containing FASTA sequence files.
  • filtered_dict.pickle (in oc23 only): A Python pickle file with preprocessed or filtered data.
  • msas: Multiple Sequence Alignments (MSAs) in .pkl format used as input for modeling.
  • pdbs: PDB structure files related to the datasets.

tar --use-compress-program=unzstd -xvf generated_models.tar.zst

└── generated_models
   ├── oc23
   │   ├── afsample2
   │   ├── SPEACH_AF
   │   ├── ...
   └── tp16
       ├── afsample2
       ├── SPEACH_AF
       ├── ...

Citation

If you use this dataset, please cite the associated publications.