Update README.md

README.md

@@ -992,60 +992,16 @@ size_categories:
 ---
 # PISCES-CulledPDB
 
-Curated protein chain tables from PISCES/CullPDB: one row per chain with sequence and metadata.
-Use the **subset dropdown** in the Hugging Face Data Viewer to switch between the main table and 242 curation subsets.
+Curated protein chain tables from PISCES/CullPDB: one row per chain with sequence and metadata. Use the **subset dropdown** in the [Hugging Face Data Viewer](https://huggingface.co/datasets/PRMegathon26/PISCES-CulledPDB) to switch between the main table and 242 curation subsets.
 
+## Dataset description
 
+The **PISCES dataset** provides curated sets of protein sequences from the Protein Data Bank (PDB) based on sequence identity and structural quality criteria. PISCES yields **non-redundant subsets of protein chains** by applying filters such as sequence identity, experimental resolution, R-factor, chain length, and experimental method (e.g., X-ray, NMR, cryo-EM). The goal is to maximize structural reliability while minimizing sequence redundancy. Unlike culling tools that rely on BLAST or global alignments, PISCES uses **PSI-BLAST** for position-specific scoring matrices, improving detection of homologs below 40% sequence identity.
 
----
-license: apache-2.0
-task_categories:
-- protein-structure-prediction
-- bioinformatics
-language:
-- en
-pretty_name: PISCES Culled Protein Chains
-size_categories:
-- 100K<n<1M
----
-
-# Dataset Card for PISCES Protein Sequence Cull Sets
-
-## Dataset Details
-
-### Dataset Description
-
-The **PISCES dataset** provides curated sets of protein sequences from the Protein Data Bank (PDB) based on user-defined sequence identity and structural quality criteria.
-
-PISCES enables users to obtain **non-redundant subsets of protein chains** by applying filters such as:
-
-- Sequence identity
-- Experimental resolution
-- R-factor
-- Chain length
-- Experimental method (e.g., X-ray, NMR, cryo-EM)
-
-The goal of PISCES is to **maximize structural reliability while minimizing sequence redundancy**.
-
-Unlike other culling tools that rely on BLAST or global alignments, PISCES uses **PSI-BLAST** to generate position-specific scoring matrices. This enables more accurate detection of homologous relationships, particularly **below 40% sequence identity**.
-
----
-
-## Dataset Sources
-
-**Original Server:**  
-PISCES
-
-**Primary Reference:**  
-Wang, G., & Dunbrack, R. L., Jr. (2003). *Bioinformatics*, 19(12), 1589–1591.
+## Dataset sources
 
-**Availability:**  
-https://dunbrack.fccc.edu/pisces/
-
-**Contact:**  
-rldunbrack@fccc.edu
-
----
+- **Server:** [PISCES](https://dunbrack.fccc.edu/pisces/)
+- **Reference:** Wang, G., & Dunbrack, R. L. Jr. (2003). *Bioinformatics* 19(12), 1589–1591.
 
 ## Citation
 
@@ -1060,53 +1016,49 @@ rldunbrack@fccc.edu
   year={2003},
   publisher={Oxford University Press}
 }
-dditional Attribution
-
-Recurated in Hugging Face by: Akshaya Narayanasamy (akshayanarayanasamy[at]gmail.com)
-
-License: Apache-2.0
-
-Primary Domain: Structural Bioinformatics / Computational Biology
+```
 
-Uses
+*Recurated for Hugging Face by Akshaya Narayanasamy akshayanarayanasamy[at]gmail.com.*
 
-This dataset can be used for:
+## Uses
 
-Generating non-redundant protein chain datasets for machine learning and statistical analysis
+- Non-redundant protein chain datasets for ML and statistical analysis  
+- Benchmarking protein structure prediction or homology modeling  
+- Studying evolutionary relationships at chosen sequence identity thresholds  
+- High-quality training sets filtered by resolution and R-factor  
+- Structure-based ML datasets for protein modeling  
 
-Benchmarking protein structure prediction or homology modeling algorithms
+## Dataset structure
 
-Studying evolutionary relationships using sequence identity thresholds
+| Item | Description |
+|------|-------------|
+| **Main CSV** | `curated_csv/cullpdb_combined_chains.csv` — full chain table |
+| **Subsets** | `curated_csv/subsets/*.csv` — 242 files (same columns) |
+| **Index** | `curated_csv/cullpdb_list_fasta_index.csv` |
 
-Designing high-quality training datasets filtered by structural resolution and reliability
+Subset paths: `curated_csv/dataset_metadata.json` (keys `data_paths`, `subset_paths`).
 
-Building structure-based machine learning datasets for protein modeling
+### Columns (chain CSVs)
 
-Dataset Structure
+| Column | Description |
+|--------|-------------|
+| **pdb_chain** | PDB chain ID (e.g. 1ABC_A) |
+| **pdb** | PDB ID (first 4 chars) |
+| **chain** | Chain ID |
+| **sequence** | Amino acid sequence (one-letter) |
+| **len** | Sequence length |
+| **method** | Experimental method (e.g. XRAY, NMR) |
+| **resolution** | Resolution in Å |
+| **rfac** | R-factor |
+| **freerfac** | Free R-factor |
+| **pc** | Sequence identity cutoff % for this subset |
+| **no_breaks** | Whether chain has no breaks (yes/no) |
+| **R** | R-factor cutoff for this subset |
+| **source_list** | Subset list basename (curation parameters) |
 
-Each record corresponds to one protein chain.
-| Field       | Description                                         |
-| ----------- | --------------------------------------------------- |
-| PDB         | Protein Data Bank identifier                        |
-| Chain       | Chain identifier within the PDB structure           |
-| Sequence    | Amino acid sequence of the chain                    |
-| Len         | Sequence length                                     |
-| Method      | Experimental method used to determine the structure |
-| Rfac        | Crystallographic R-factor                           |
-| Freerfac    | Free R-factor                                       |
-| Pc          | Maximum pairwise sequence identity threshold        |
-| Break       | Indicates whether the chain contains breaks         |
-| R           | Resolution of the structure                         |
-| Source_list | Original PISCES list source                         |
+## Usage
 
-Sequences are sorted by structural quality metrics such as resolution and R-factor and then culled based on mutual sequence identity thresholds to ensure non-redundancy.
-| Item         | Description                                                  |
-| ------------ | ------------------------------------------------------------ |
-| **Main CSV** | `curated_csv/cullpdb_combined_chains.csv` — full chain table |
-| **Subsets**  | `curated_csv/subsets/*.csv` — 242 files (same columns)       |
-| **Index**    | `curated_csv/cullpdb_list_fasta_index.csv`                   |
-Full list of subset paths:
-curated_csv/dataset_metadata.json (keys data_paths, subset_paths).
+```python
 from huggingface_hub import hf_hub_download
 import pandas as pd
 
@@ -1115,34 +1067,24 @@ path = hf_hub_download(
     filename="curated_csv/cullpdb_combined_chains.csv",
     repo_type="dataset"
 )
-
 df = pd.read_csv(path)
-File Naming Convention
-
-Each dataset file follows the naming pattern:
-cullpdb_pc{pc}_res{res_min}-{res_max}[_noBrks]_len40-10000_R{R}_{methods}_d2026_01_26_chains{N}
-Parameters
-
-pc – Percent sequence identity cutoff
-Examples: 15, 20, 25, …, 95
-
-res – Resolution range (Å)
-Examples: 0.0-1.0, 0.0-2.5, 0.0-5.0
-
-noBrks – Optional flag indicating chains with breaks are excluded
-
-R – R-factor cutoff
-Possible values: 0.2, 0.25, 0.3, 1.0
+```
 
-methods – Experimental structure determination methods
+## File naming convention
 
-Possible values:
+Subset filenames follow:
 
-Xray
+`cullpdb_pc{pc}_res{res_min}-{res_max}[_noBrks]_len40-10000_R{R}_{methods}_d2026_01_26_chains{N}.csv`
 
-Xray+EM
+| Parameter | Meaning |
+|-----------|---------|
+| **pc** | Percent sequence identity cutoff (15, 20, …, 95) |
+| **res** | Resolution range in Å (e.g. 0.0-1.0, 0.0-2.5) |
+| **noBrks** | Optional: exclude chains with breaks |
+| **R** | R-factor cutoff (0.2, 0.25, 0.3, 1.0) |
+| **methods** | Xray, Xray+EM, or Xray+Nmr+EM |
+| **N** | Number of chains in the list |
 
-Xray+Nmr+EM
+## License
 
-N – Total number of chains in the list
-```
+Apache-2.0