README.md

---
license: apache-2.0
base_model: zhihan1996/DNABERT-2-117M
tags:
  - genomics
  - variant-classification
  - dnabert2
  - clinvar
  - bioinformatics
  - bert
pipeline_tag: text-classification
language:
  - dna
---

# DNABERT-2 — ClinVar Variant Pathogenicity Classifier

Fine-tuned from [zhihan1996/DNABERT-2-117M](https://huggingface.co/zhihan1996/DNABERT-2-117M)
on **ClinVar high-confidence (≥2 review stars) single nucleotide variants (SNVs)**
for binary pathogenicity classification.

| Label | ID | Meaning |
|---|---|---|
| Benign | 0 | Benign / Likely benign |
| Pathogenic | 1 | Pathogenic / Likely pathogenic |

## Test-set metrics

Gene-level 80/10/10 split on GRCh38 SNVs (no gene appears in both train and test).
Test set is 16.3% pathogenic — class-weighted loss used during training.

| Metric | Value |
|---|---|
| **AUROC** | **0.8342589406111742** |
| AUPRC | 0.6144893948752504 |
| F1 (pathogenic class) | 0.5413069162955557 |
| MCC | 0.44952950405729725 |
| Accuracy | 0.8467494610269335 |

Cancer gene panel (BRCA1/2, TP53, MLH1/2, MSH2/6, ATM, CHEK2, PALB2, APC, PTEN, RAD51C):
**AUROC 0.819 · AUPRC 0.733 · F1 0.629**

## Input format

A **512 bp DNA string** (characters A/C/G/T only), centered on the variant position
with the **ALT allele introduced at index 255** (0-based centre of the window).
Extracted from the GRCh38 reference genome using a ±256 bp window.

## Usage

```python
from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

tok = AutoTokenizer.from_pretrained("whitedevil0089devil/dnabert2-clinvar-pathogenicity", trust_remote_code=True)
model = AutoModelForSequenceClassification.from_pretrained(
    "whitedevil0089devil/dnabert2-clinvar-pathogenicity", trust_remote_code=True)
model.eval()

# sequence = 512 bp window from GRCh38, ALT allele injected at position 255
sequence = "ACGT" * 128   # replace with your real sequence

inputs = tok(sequence, return_tensors="pt", truncation=True, max_length=512)
with torch.no_grad():
    logits = model(**inputs).logits

prob_pathogenic = torch.softmax(logits, dim=-1)[0, 1].item()
label = "Pathogenic" if prob_pathogenic >= 0.5 else "Benign"
print(f"Pathogenicity probability: {prob_pathogenic:.4f}  →  {label}")
```

## Confidence tiers (used in the demo app)

| Score | Tier | Suggested action |
|---|---|---|
| 0.00–0.20 | Likely Benign | Routine monitoring |
| 0.20–0.40 | Possibly Benign | Note in record |
| 0.40–0.60 | Uncertain | Flag for expert review |
| 0.60–0.80 | Possibly Pathogenic | Prioritize functional validation |
| 0.80–1.00 | Likely Pathogenic | Urgent — genetic counseling |

## Data and training

- **Source:** ClinVar `variant_summary.txt` (GRCh38, SNVs, ≥2 review stars, 2026)
- **Labels:** Pathogenic + Likely pathogenic → 1 · Benign + Likely benign → 0 · VUS dropped
- **Split:** by `GeneSymbol` (prevents gene-level leakage)
- **Context:** 512 bp window from GRCh38 reference, ALT allele injected
- **Imbalance:** class-weighted CrossEntropyLoss (16.3% pathogenic)
- **Optimizer:** AdamW · LR 3e-5 · warmup 6% · cosine decay · 5 epochs · batch 64
- **Hardware:** NVIDIA RTX PRO 6000 Blackwell (96 GB) · fp16

## Limitations

- SNVs only — insertions, deletions, structural variants not supported
- Trained on germline ClinVar variants — not validated on somatic mutations
- Gene-level split means genes absent from ClinVar are unseen during training
- AUROC of 0.83 is good for ranking; the model should not be used as a
  sole clinical decision-making tool

## Citation

```bibtex
@inproceedings{zhou2024dnabert2,
  title={DNABERT-2: Efficient Foundation Model and Benchmark for Multi-Species Genome},
  author={Zhou, Zhihan and Ji, Yanrong and Li, Weijian and Dutta, Pratik and
           Davuluri, Ramana and Liu, Han},
  booktitle={ICLR},
  year={2024}
}
```
ClinVar: Landrum et al., Nucleic Acids Research 2020.