Dataset Viewer
Duplicate
The dataset viewer is not available for this split.
Job manager crashed while running this job (missing heartbeats).
Error code:   JobManagerCrashedError

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

STXBP1 Base Editing Parameter Sweep

654 million parameter combinations analyzed for 170 pathogenic STXBP1 variants

Overview

This dataset contains exhaustive parameter optimization results for base editing guide design targeting STXBP1 mutations. It identifies which variants can be studied in wild-type mouse models versus those requiring humanized mice.

πŸ† Perfect Score Variants (100/100)

Variant cDNA Editor WT Mouse Compatible
E53X c.157G>T Prime βœ… YES
A297S c.889G>T Prime βœ… YES
E302X c.904G>T Prime βœ… YES
M443R c.1328T>G Prime βœ… YES
L446F c.1336C>T CBE βœ… YES

Dataset Statistics

Metric Value
Total analyses 654,595,200
Variants 170 SNVs
Parameter combos/variant 3,850,560
Compatible variants ~95
Compute time 35 minutes
Hardware NVIDIA GH200 (ARM64 + H100)

Files

  • stxbp1_ULTIMATE_best.json - Optimal parameters for each variant (recommended)
  • stxbp1_ULTIMATE_sweep.csv.gz - All 654M analyses (~5GB compressed)

Parameter Space

Parameter Range Values
Scan radius 10-200 bp 191
Context radius 5-100 bp 96
Edit window (1,1) to (20,20) 210

Scoring System (0-100)

  • Context identity (40 pts): Human-mouse sequence conservation
  • Guide quality (30 pts): Protospacer mismatch penalty
  • Edit window (20 pts): Target base in optimal editing positions
  • WT compatible (10 pts): Can use wild-type mouse

Editor Types

Editor Change Count Example
ABE A→G 32 ABE8e
CBE C→T 50 BE4max
Prime Any 82 PE2/PE3

Usage

import json

# Load optimal configurations
with open('stxbp1_ULTIMATE_best.json') as f:
    best = json.load(f)

# Get K196X optimal parameters
k196x = best['K196X']
print(f"Score: {k196x['score']}")
print(f"Editor: {k196x['edit_type']}")
print(f"Compatible: {k196x['compatible']}")

Query the Full CSV

import duckdb

# Find all parameter combos for K196X scoring > 90
results = duckdb.query("""
    SELECT * FROM 'stxbp1_ULTIMATE_sweep.csv.gz'
    WHERE label = 'K196X' AND score > 90
    ORDER BY score DESC
    LIMIT 100
""").df()

Reference Sequences

  • Human: NM_001032221.6 (1,785 bp, 594 aa)
  • Mouse: NM_009295 (1,785 bp, 594 aa)
  • Overall identity: 90.9%

Interactive Tool

Test individual variants: STXBP1 Base Editing Verifier

Citation

@dataset{stxbp1_base_editing_sweep_2026,
  author = {Freygang, Adam},
  title = {STXBP1 Base Editing Mouse Model Compatibility Parameter Sweep},
  year = {2026},
  publisher = {HuggingFace},
  url = {https://huggingface.co/datasets/SkyWhal3/stxbp1-base-editing-sweep}
}

License

MIT

Acknowledgments

  • STXBP1 Foundation
  • Lambda Cloud (compute)
  • STXBP1-ARIA Project

For research use. Not medical advice.

Downloads last month
14