comp-genomics-consortium/synthetic-protein-folding-matrices

Dataset Card for CGSC Synthetic Protein Folding Matrices

Dataset Summary

This is the secondary repository for the CGSC, strictly dedicated to archiving high-resolution, uncompressed molecular dynamics (MD) trajectories. The dataset contains continuous temporal matrices representing synthetic protein folding simulations at sub-angstrom resolution.

Simulating atomic interactions over time generates colossal amounts of data. To preserve the micro-fluctuations and raw temporal states required for training next-generation folding algorithms, these trajectory matrices are stored as raw, uncompressed binary dumps rather than standard .pdb or .xtc formats. This repository is intended to stress-test high-throughput tensor ingestion and structural prediction models.

Supported Tasks and Leaderboards

• structural-trajectory-prediction: Utilizing uncompressed temporal data to forecast final protein states.
• molecular-dynamics-ingestion: Stress-testing data pipelines against massive, unformatted multi-dimensional arrays.

Languages

There is no natural language present in this dataset. Internal headers and metadata are formatted in English (en), but the payloads are entirely numerical and binary.

Dataset Structure

Data Instances

Files in this repository do not follow standard structured tabular formats. They are massive coordinate dumps.

{
  "simulation_id": "cgsc-fold-alpha-110",
  "simulation_timestamp_utc": "2026-05-12T09:45:00Z",
  "matrix_type": "temporal_coordinate_dump",
  "compression_state": "raw_uncompressed_trajectory",
  "payload_reference": "matrices/fold-110-trajectory.bin",
  "atomic_resolution": "sub-angstrom"
}

Data Fields

• simulation_id: Unique identifier for the specific folding simulation.
• simulation_timestamp_utc: Time the simulation was completed.
• matrix_type: Categorization of the array (e.g., temporal_coordinate_dump).
• compression_state: Maintained as raw_uncompressed_trajectory to explain the massive file sizes.
• payload_reference: Direct pointer to the heavy binary matrices in LFS.
• atomic_resolution: Indicates the fidelity of the raw dump.

Dataset Creation

Curation Rationale

Traditional structural biology databases utilize heavy compression, which often obscures micro-states in molecular dynamics. The CGSC archives these uncompressed matrices to provide a raw, high-entropy playground for advanced model training, at the cost of immense storage requirements.

Source Data

All trajectories are strictly synthetic, computationally generated on CGSC's internal GPU clusters. No organic biological samples were used.

Disclaimers

Due to the uncompressed temporal nature of molecular dynamics, these matrices are exceptionally large. Downloads are strictly recommended only for research institutions with petabyte-scale storage capabilities and specialized tensor-parsing infrastructure.