Portfolio

Clinical and microbial genomics projects. The clinical pipelines are benchmarked against public truth sets (GIAB) and generate a clinical-style PDF report. All data are public or simulated, never patient data.

Clinical Genomics

Germline WGS Benchmarked on GIAB In progress

FASTQ to ACMG-classified variants, benchmarked on seven GIAB genomes (HG001–HG007), with GATK and DeepVariant compared side by side.

Nextflow · Docker · BWA-MEM2 · GATK · DeepVariant · hap.py · VEP · Exomiser

Microbial Genomics

Erythromycin Resistance in S. aureus: DNABERT-2 vs XGBoost In progress

Predicting erythromycin resistance in Staphylococcus aureus with a fine-tuned DNA language model compared with XGBoost on gene features.

Python · uv · PyTorch · Hugging Face Transformers · DNABERT-2 · LoRA · XGBoost · AMRFinderPlus · NCBI Datasets