dnabert2

🧬 GenomeIQ — Explainable AI for DNA disease classification. Predicts Cancer, Diabetes, Alzheimer's, or Normal markers from raw DNA sequences using TF-IDF k-mer ensemble + fine-tuned DNABERT-2 transformer. Features saliency maps, OOD detection, BLAST-lite similarity, mutation analysis, and a Gemini-powered RAG copilot.

machine-learning bioinformatics deep-learning genomics transformers disease-classification explainable-ai dnabert2

• Updated May 26, 2026
• Python

Fine-tuning DNABERT-2 to classify normal vs tumour colorectal gene enhancers from DNA sequence alone (MSc thesis). PR-AUC 0.758 · arXiv:2509.25274

machine-learning bioinformatics deep-learning genomics scikit-learn transformers computational-biology pytorch enhancers optuna dna-sequence hugging-face colorectal-cancer genomic-language-model dnabert2

• Updated Jul 24, 2026
• Jupyter Notebook