I'm a bioinformatician in Toronto with a background in health research.

For four years I managed and analyzed data for the MASHAD cohort, a cardiovascular study of about 9,000 participants. I kept the participant data in PostgreSQL, derived nutrient patterns in Python using PCA, compared them between people with and without cardiovascular disease, and co-authored several papers along the way.

I then did a Master of Bioinformatics at the University of Guelph. My project compared the leaf microbiomes of herbarium specimens (some from 1938) with field-collected plants of a wild tomato relative. I built the workflow in Nextflow on the Digital Research Alliance of Canada HPC (SLURM), with tools for degraded DNA (leeHom, Samtools, Picard, BBMap), Kraken2 for classification and QIIME2 for diversity analysis.

Right drug, right patient: How far can we trust a cancer trial? Subgroup findings in a simulated lung cancer trial (CDISC SDTM/ADaM, admiral, tern, bonsaiforest) and a real-world chemotherapy cohort in the OMOP CDM.

Predicting cancer origin: Can gene expression and mutations tell us where a metastatic tumour started? Uses data from 570 BC Cancer patients, with scikit-learn, XGBoost and SHAP, and a Docker + Nextflow pipeline.

Python · R · SQL (PostgreSQL) · Bash · Nextflow · Docker · Git · HPC/SLURM · Quarto