Portfolio
Selected projects spanning machine learning, reproducible pipelines, and web tools — from coursework, freelance work, and independent learning.
Machine Learning & Predictive Models
Models and tools applying machine learning to biological prediction problems.
Vaccine Design
Vaccine Epitope Prediction Dashboard
A deployed Streamlit app predicting candidate vaccine epitopes from a protein sequence using MHCflurry, ranking fragments as Strong/Weak/Non-Binder with interactive charts and CSV export.
ML Classifier
Breast Cancer Subtype Classification
Random Forest classifier predicting malignant vs. benign tumors from gene expression features — 96.5% accuracy on the Wisconsin dataset.
Thesis
MS Thesis: TGF-β Dysregulation in HCC
Multi-omics analysis of TGF-β signaling dysregulation in hepatocellular carcinoma using TCGA and GEO cohorts. Full results published on thesis submission.
Reproducible Pipelines
Containerized, reproducible analysis pipelines built with Nextflow and Docker — designed to run identically across environments.
RNA-SEQ
RNA-Seq Differential Expression — C. elegans
End-to-end pipeline identifying sex-specific gene expression differences. 947 significant DEGs found, with volcano plot and heatmap visualization via DESeq2.
PIPELINE
RNA-seq Nextflow + Docker Pipeline
A reproducible RNA-seq pipeline (FastQC → Trimmomatic → STAR → featureCounts → MultiQC), fully containerized so it runs identically anywhere.
PIPELINE
Metagenomics Nextflow + Docker Pipeline
Containerized metagenomics pipeline for taxonomic classification. QC and trimming stages fully built and tested; classification stage requires an external database too large to bundle.
Applications
Web applications built for biological data analysis and visualization.
WEB APP
ExpressionProfiler
A Flask web app for differential gene expression analysis, built as my BS final year project. Won 1st place at the COMSATS bioinformatics expo. Code is private per department policy.