| --- |
| title: FunGO |
| emoji: 🧬 |
| colorFrom: blue |
| colorTo: green |
| sdk: gradio |
| sdk_version: 4.44.0 |
| app_file: app.py |
| pinned: false |
| license: mit |
| short_description: Protein Function Prediction using ESM2 + XGBoost |
| --- |
| |
| # FunGO — Protein Function Prediction |
|
|
| **Beyond Prediction — Understanding Function.** |
|
|
| FunGO predicts Gene Ontology (GO) terms for protein sequences using: |
| - **ESM2-t36-3B** — protein language model embeddings (layers 30–35) |
| - **XGBoost classifiers** — 4,133 GO-term specific models |
| - **Evidence-tiered filtering** — Strong / Moderate / Indicative |
|
|
| ## Evidence Tiers |
|
|
| | Tier | IA Weight | Confidence | Description | |
| |------|-----------|------------|-------------| |
| | Strong Evidence | > 5.0 | ≥ 0.30 | Highly specific GO term | |
| | Moderate Evidence | > 2.0 | ≥ 0.50 | Moderately specific term | |
| | Indicative | > 1.0 | ≥ 0.65 | Lower specificity, high confidence | |
|
|
| ## Ontologies Covered |
|
|
| - **MFO** — Molecular Function |
| - **BPO** — Biological Process |
| - **CCO** — Cellular Component |
|
|
| ## Development Team |
|
|
| - **Dr. Beenish Maqsood** — Principal Investigator, School of Biochemistry and Biotechnology, University of the Punjab |
| - **Dr. Naeem Mahmood** — Co-Supervisor, School of Biochemistry and Biotechnology, University of the Punjab |
| - **Muteeba Azhar** — Lead Developer, MS Researcher, University of the Punjab |
|
|
