Regulatory pharmacy
Myanmar's FDA
I reviewed medicines after approval and manufacture. The work was important, but reactive: I was checking outcomes rather than the biological causes behind them.
Research and career motivation
My goal is to turn patient omics data into tools clinicians can use.
I want to build biomarker panels, risk-stratification models, and treatment-response predictors from proteomic, transcriptomic, and genomic data — not as abstract models, but as practical support for clinical decisions. For me, precision medicine starts with one human question: what will help this patient?
Origin
From checking outcomes to predicting response.
I trained as a pharmacist. At Myanmar's FDA, my work focused on reviewing medicines after they had already been approved and manufactured. That work taught me the importance of safety, regulation, and clinical responsibility — but it was mostly reactive.
My MSc thesis changed the question. Using label-free LC-MS/MS proteomics, I looked for serum proteins that could predict chemotherapy response in non-small cell lung cancer. For the first time, I was not asking, "Is this drug safe?" I was asking, "Will this specific patient respond to this specific treatment?"
Same question, different molecular layers
Proteome, exome, and transcriptome projects all pointed toward one direction.
LC-MS/MS proteomics
Prince of Songkla University
My MSc thesis used label-free serum proteomics to identify proteins associated with chemotherapy response in non-small cell lung cancer.
Whole-exome sequencing
Chiang Mai University
I built BWA, GATK, ANNOVAR, and VEP pipelines to identify pharmacogenomic variants relevant to osteosarcoma treatment.
RNA-seq transcriptomics
National Taiwan University
I applied limma and edgeR differential-expression analysis to colorectal and breast cancer cohorts, testing whether transcriptomic signatures could distinguish disease stages.
OLINK proteomics and machine learning
IGTP, Barcelona
I trained Random Forest models on 1,395 proteins from 450 Long COVID patients to stratify outcomes, reaching more than 85% cross-validated accuracy.
Why computational methods
Clinical prediction needs scale.
Wet-lab work can answer one protein or one gene at a time. Computational biology lets me ask the same clinical question across thousands of features, hundreds of patients, and the time constraints of a real treatment decision.
That is why my path moved from bench-based proteomics toward Python- and R-based pipeline development, statistical modeling, machine learning, and multi-omics integration.
Why this work fits me
Pharmacy gives me the clinical frame. Computation gives me the scale.
My pharmacy background gives me working knowledge of drug mechanisms, regulatory constraints, and clinical outcomes. My omics and computational experience gives me the tools to model those outcomes from patient-level molecular data.
This combination is the basis of my current specialization: computational biology applied to precision medicine.
What I want next
A role where predictive models move beyond the manuscript.
I want to build predictive models from patient omics data, validate them against clinical outcomes, and see them used in a real clinical or translational pipeline. Not authorship for its own sake: application. A model that stays in a manuscript has not answered the original question. A model that changes a treatment decision has.