DeepVRegulome is an end‑to‑end framework for predicting the functional impact of small somatic variants in non‑coding regulatory regions (splice sites and transcription‑factor‑binding sites) using fine‑tuned DNABERT models.

Reference-free, privacy-preserving wastewater pathogen surveillance for federated hospital networks. DNABERT-2 embeddings + Transformer-encoder VAE detect emerging anomaly clusters in unclassified reads. AIxBio Hackathon Track 2, Apart Research.

Assessing LLMs—AAM, DNABERT, DNABERT-2, GROVER—to improve the prediction of COVID-19 status ("Positive" or "Not detected") using microbiome data

Code and datasets for studying how negative dataset selection impacts ML-based promoter predictors across multiple bacterial species (Random Forest, CNN, DNABERT, Nucleotide Transformer).

Mosquito Virome Emergence Framework — companion code for Hoyos-López (2026), Briefings in Bioinformatics

Predict antomicrobial resistance (AMR) for a pathogen (Staphylococcus aureus) against an antibiotic (Cefoxitin) using encoder-only transformer given the genomic sequence of one of its genes (gene pbp4) as input.

My Implementation of the DNABERT papers.

From Raw DNA to Real Action. An AI-powered platform that translates raw river eDNA datasets into instant, plain-English biodiversity intelligence and automatically alerts government agencies of ecological risks in minutes instead of years.