Datasets:
Upload folder using huggingface_hub
Browse files- README.md +74 -0
- glof_dataset.tsv +0 -0
README.md
ADDED
|
@@ -0,0 +1,74 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: cc-by-4.0
|
| 3 |
+
task_categories:
|
| 4 |
+
- tabular-classification
|
| 5 |
+
language:
|
| 6 |
+
- en
|
| 7 |
+
tags:
|
| 8 |
+
- biology
|
| 9 |
+
- genomics
|
| 10 |
+
- clinical-genetics
|
| 11 |
+
- missense-variants
|
| 12 |
+
- loss-of-function
|
| 13 |
+
- gain-of-function
|
| 14 |
+
- variant-effect-prediction
|
| 15 |
+
- benchmark
|
| 16 |
+
- precision-medicine
|
| 17 |
+
size_categories:
|
| 18 |
+
- 100K<n<1M
|
| 19 |
+
---
|
| 20 |
+
|
| 21 |
+
# GLOF: A large-scale expert-curated benchmark of missense variant functional effects
|
| 22 |
+
|
| 23 |
+
GLOF (Gain and Loss Of Function) is a benchmark dataset of **112,399 missense variants** across **2,809 human genes**, each classified as **LOF** (loss-of-function), **GOF** (gain-of-function), or **Neutral** by board-certified clinical geneticists.
|
| 24 |
+
|
| 25 |
+
## Dataset Description
|
| 26 |
+
|
| 27 |
+
The dataset was curated at [Mendelics Analise Genomica](https://www.mendelics.com/), one of Latin America's largest clinical genomics laboratories. The annotation process integrated ClinVar pathogenicity classifications, published functional studies, established gene-disease relationships, and expert clinical judgment following ACMG/AMP guidelines.
|
| 28 |
+
|
| 29 |
+
- **Pathogenic variants** were sourced from ClinVar (July 2023 release) and cross-referenced against the March 2026 release; variants with reclassified or conflicting evidence were excluded.
|
| 30 |
+
- **Neutral variants** were drawn from gnomAD v3.1 and validated against v4.1 allele frequencies, selecting missense variants with AF > 1%.
|
| 31 |
+
|
| 32 |
+
## Dataset Schema
|
| 33 |
+
|
| 34 |
+
| Field | Type | Description |
|
| 35 |
+
|-------|------|-------------|
|
| 36 |
+
| VARIANTKEY | String | Unique variant identifier: `chr-position-ref-alt` (GRCh38) |
|
| 37 |
+
| LABEL | String | Functional classification: `Neutral`, `LOF`, or `GOF` |
|
| 38 |
+
| ENSG | String | Ensembl gene identifier |
|
| 39 |
+
| GENE_SYMBOL | String | HGNC gene symbol |
|
| 40 |
+
| AA_POSITION | Integer | Amino acid substitution position in the canonical protein |
|
| 41 |
+
| PROTEIN_REF | Character | Reference (wild-type) amino acid |
|
| 42 |
+
| PROTEIN_ALT | Character | Alternate (mutant) amino acid |
|
| 43 |
+
|
| 44 |
+
## Class Distribution
|
| 45 |
+
|
| 46 |
+
| Class | Variants | Percentage | Genes |
|
| 47 |
+
|-------|----------|------------|-------|
|
| 48 |
+
| Neutral | 83,902 | 74.6% | 2,749 |
|
| 49 |
+
| LOF | 25,368 | 22.6% | 2,020 |
|
| 50 |
+
| GOF | 3,129 | 2.8% | 260 |
|
| 51 |
+
|
| 52 |
+
## Usage
|
| 53 |
+
|
| 54 |
+
```python
|
| 55 |
+
from datasets import load_dataset
|
| 56 |
+
|
| 57 |
+
dataset = load_dataset("victormaricato/glof")
|
| 58 |
+
```
|
| 59 |
+
|
| 60 |
+
## Citation
|
| 61 |
+
|
| 62 |
+
If you use this dataset, please cite:
|
| 63 |
+
|
| 64 |
+
```bibtex
|
| 65 |
+
@article{maricato2026glof,
|
| 66 |
+
title={GLOF: A large-scale expert-curated benchmark dataset of gain-of-function and loss-of-function missense variants},
|
| 67 |
+
author={Maricato, Victor and Schlesinger, David and de Souza Moura, Pedro Nuno},
|
| 68 |
+
year={2026}
|
| 69 |
+
}
|
| 70 |
+
```
|
| 71 |
+
|
| 72 |
+
## License
|
| 73 |
+
|
| 74 |
+
This dataset is released under the [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/) license.
|
glof_dataset.tsv
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|