Projects
Software tools for proteomics research and data analysis
NULISA Analysis Software (NAS)
Lead Developer & Architect
A web application for analyzing proteomics data from the NULISA platform. NAS gives researchers tools for quality control, statistical modeling, and interactive visualization. It processes thousands of protein targets across hundreds of samples, with built-in quality metrics and differential expression analysis.
Technology Stack
Architecture & Major Modules
- 55 R modules organized hierarchically: Project Management (6), Quality Control (8), Statistical Analysis (5), Data Visualization (6), Normalization (4), Cloud Infrastructure (4), and supporting utilities
- Statistical modeling framework supporting linear models and linear mixed-effects models for differential expression and longitudinal study designs
- Multi-dimensional QC system with detectability analysis, precision metrics (intra/inter-plate CV), internal control monitoring, and plate layout validation
- Interactive visualization suite featuring customizable heatmaps (iheatmapr), PCA plots, boxplots, and calibration curves for absolute quantification
- Cloud-native architecture with Azure AD authentication, Key Vault secrets management, Blob Storage persistence, and PostgreSQL telemetry
Technical Achievements
- Built reactive state management across 55 interconnected modules with real-time data updates throughout the app
- Set up automated CI/CD pipelines with quality gates (linting, unit tests, E2E automation) for dev and production environments
- Developed custom forks of R visualization packages (iheatmapr, PCAtools) to extend capabilities for NULISA-specific analytical requirements
- Implemented full-stack testing strategy using shinytest2 for module testing and Cypress for end-to-end UI workflow automation
- Built telemetry infrastructure tracking user interactions across all modules for data-driven product improvements
- Built data validation and sanitization pipeline for covariate management and data integrity
Other Projects
NULISAseqR Package
Core contributor to the NULISAseqR R package, developing essential functions for NULISA data analysis. Built multi-plate merging pipeline and ETL infrastructure, adding project-level analysis capabilities that didn't exist before.
PG&E Data Visualizer
Energy analytics platform built with R Shiny featuring four analysis engines for smart meter data. Implemented multiple anomaly detection algorithms (IQR, Z-Score, STL, Moving Average), k-means pattern clustering, and automated Excel reporting with production CI/CD deployment.