---
license: cc-by-4.0
task_categories:
  - other
tags:
  - biology
  - protein
  - structure
  - PDB
  - PISCES
  - CullPDB
  - sequence
  - curation
language: en
size_categories:
  - n>1M
dataset_info:
  config_name: default
  features:
    - name: pdb_chain
      dtype: string
    - name: pdb
      dtype: string
    - name: chain
      dtype: string
    - name: sequence
      dtype: string
    - name: len
      dtype: int64
    - name: method
      dtype: string
    - name: resolution
      dtype: float64
    - name: rfac
      dtype: string
    - name: freerfac
      dtype: string
    - name: pc
      dtype: float64
    - name: no_breaks
      dtype: string
    - name: R
      dtype: float64
    - name: source_list
      dtype: string
  splits: null
  num_rows: 4540884
# Main and index paths; full list of subset paths is in dataset_metadata.json
data_paths:
  main_csv: curated_csv/cullpdb_combined_chains.csv
  index_csv: curated_csv/cullpdb_list_fasta_index.csv
  full_list_csv: curated_csv/cullpdb_full_compiled_list.csv
  subsets_dir: curated_csv/subsets
  metadata: curated_csv/dataset_metadata.json
---

# PISCES-CulledPDB — Curated chain CSVs

Curated protein chain tables derived from PISCES/CullPDB-style lists and FASTA files. One row per chain with sequence, metadata, and curation parameters.

**Dataset:** [PRMegathon26/PISCES-CulledPDB](https://huggingface.co/datasets/PRMegathon26/PISCES-CulledPDB)

## Dataset Summary

| Item | Description |
|------|-------------|
| **Main CSV** | `curated_csv/cullpdb_combined_chains.csv` — single table with **4,540,884** chains (one row per chain). |
| **Subset CSVs** | `curated_csv/subsets/*.csv` — **242** files, one per curation subset (same columns as main). |
| **Index** | `curated_csv/cullpdb_list_fasta_index.csv` — **242** rows; maps subset basenames to list/FASTA paths and parameters. |

## Data paths (metadata)

**Main and index files:**
- `curated_csv/cullpdb_combined_chains.csv` — master chain table
- `curated_csv/cullpdb_list_fasta_index.csv` — subset index
- `curated_csv/cullpdb_full_compiled_list.csv` — full compiled list
- `curated_csv/subsets/` — directory containing one CSV per subset (**242** files)

**Subset paths:** The full list of subset CSV paths is in **`curated_csv/dataset_metadata.json`** (keys `data_paths` and `subset_paths`). Use it for programmatic listing or with `hf_hub_download(..., filename=path)`.

## Main CSV & Subset CSV — Column definitions

Each chain row has the following columns:

| Column | Type | Description |
|--------|------|-------------|
| **pdb_chain** | string/number | PDB chain ID (e.g. 1ABC_A) |
| **pdb** | string/number | PDB ID (first 4 chars) |
| **chain** | string/number | Chain ID |
| **sequence** | string/number | Amino acid sequence (one-letter) |
| **len** | string/number | Sequence length |
| **method** | string/number | Experimental method (e.g. XRAY, NMR) |
| **resolution** | string/number | Resolution in Å (per structure) |
| **rfac** | string/number | R-factor |
| **freerfac** | string/number | Free R-factor |
| **pc** | string/number | Sequence identity cutoff % used for this subset |
| **no_breaks** | string/number | Whether chain has no breaks (yes/no) |
| **R** | string/number | R-factor cutoff used for this subset |
| **source_list** | string/number | Subset list basename (identifies curation parameters) |

Subset CSVs use the same schema; each file corresponds to one row in the index (one set of parameters: pc, resolution range, no_breaks, R, Nmethods).

## Index CSV — Column definitions

| Column | Description |
|--------|-------------|
| **list_basename** | Subset list basename |
| **fasta_basename** | Corresponding FASTA basename |
| **list_path** | Full path to list file |
| **fasta_path** | Full path to FASTA file |
| **n_chains** | Number of chains in this subset |
| **pc** | Sequence identity cutoff % |
| **resolution** | Resolution range (e.g. 0.0-2.0) |
| **no_breaks** | yes/no |
| **R** | R-factor cutoff |
| **Nmethods** | Experiment types (e.g. Xray, Xray+EM) |

## Subset naming

Subset list basenames follow:

`cullpdb_pc{pc}_res{min}-{max}_[noBrks_]len40-10000_R{R}_{Nmethods}_..._chains{N}`

- **pc**: sequence identity cutoff (%)
- **resolution**: Å range (e.g. 0.0–2.0)
- **noBrks** (optional): no breaks in chain
- **R**: R-factor cutoff
- **Nmethods**: Xray, Xray+EM, or Xray+Nmr+EM
- **N**: number of chains in that subset

## Usage

```python
from huggingface_hub import hf_hub_download
import pandas as pd

# Download main CSV (large)
path = hf_hub_download(repo_id="PRMegathon26/PISCES-CulledPDB", filename="curated_csv/cullpdb_combined_chains.csv", repo_type="dataset")
df = pd.read_csv(path)

# Or load a single subset
path = hf_hub_download(repo_id="PRMegathon26/PISCES-CulledPDB", filename="curated_csv/subsets/cullpdb_pc20.0_res0.0-1.0_len40-10000_R0.2_Xray_d2026_01_26_chains314.csv", repo_type="dataset")
df = pd.read_csv(path)
```

## Curation workflow

1. **Index**: List/FASTA pairs scanned → `cullpdb_list_fasta_index.csv`
2. **Build**: For each index row, list and FASTA aligned 1:1 → one chain CSV per subset + combined master CSV
3. **Filter/visualize**: Optional filtering and summary figures from the master CSV

## License

CC BY 4.0 (or as specified in the repo).