Skip to content

vd4mmind/BioTrial-Analytics

Repository files navigation

BioTrial Analytics Logo

BioTrial Analytics

Precision Clinical Trial Visualization & Multi-Omics Power Planning

React TypeScript Tailwind License


🔬 Overview

BioTrial Analytics is a high-fidelity clinical trial simulation and power planning platform. It bridges the gap between retrospective discovery (visualizing longitudinal biomarker trends) and prospective study design (calculating statistical power for high-dimensional omics).

Built for biostatisticians and clinical researchers, the platform implements the PoweREST framework for hierarchical spatial modeling and advanced multiplicity correction for proteomics and metabolomics.


🏗️ Core Pillars

1. Clinical Discovery Dashboard (Retrospective)

  • Simulated Cohorts: Generate realistic Phase IIb data for N=600 patients across three study arms (Placebo, Drug X 1mg, Drug X 2mg).
  • Longitudinal Trends: Visualize Mean ± SEM over 24 weeks with support for Log Scale and % Change from Baseline.
  • Pharmacodynamics: Automated AUC (Area Under the Curve) calculation using the trapezoidal rule to assess biomarker sustainability.
  • Distribution Analysis: Patient-level scatter plots to identify responders vs. non-responders.

2. Multi-Omics Power Designer (Prospective)

  • Proteomics: Support for ELISA, Olink, and SomaScan. Models biological variability vs. technical noise.
  • Metabolomics & Lipidomics: Plan discovery studies for Metabolon, Nightingale, and Biocrates.
  • Multiplicity Tax: Integrated Bonferroni, FDR, and FWER correction engines to handle high-plex platforms (up to 7000+ analytes).
  • Correlation Modeling: Accounts for feature-level correlation in lipidomics to optimize class-level enrichment power.

3. Single-Cell & Spatial Transcriptomics

  • scRNA-seq (Pseudobulk LMM): Models power for paired designs using pseudobulk aggregation and rare cell type detection constraints.
  • Spatial Power (PoweREST): Implements the hierarchical variance framework:
    • Variance Decomposition: Subjects ($\sigma^2_p$), Slices ($\sigma^2_s$), and Technical replicates ($\sigma^2_t$).
    • Longitudinal Gain: Models the statistical benefit of repeated measures: $G_{lon} = 1 + (T-1)\rho$.
  • Tissue Dynamics: Interactive Canvas-based simulation of tissue remodeling and immune infiltration.

📐 Statistical Framework

BioTrial Analytics isn't just a UI; it's a statistical engine.

  • Hierarchical Modeling: Uses nested variance components for spatial and single-cell designs.
  • Noise Calibration: Pre-calibrated noise profiles for major omics platforms (e.g., Olink Explore, SomaScan v4).
  • Efficiency Frontier: Automated optimization of Subject N vs. Slice Count vs. Cost to find the precision-budgeting "sweet spot."

💻 Tech Stack

  • Frontend: React 19 (Functional Components, Hooks)
  • State Management: Context API & Optimized Local State
  • Visualization: Recharts (Custom SVG labeling) & HTML5 Canvas (High-performance simulations)
  • Styling: Tailwind CSS (Mobile-first, responsive design)
  • Icons: Lucide React
  • Type Safety: Strict TypeScript 5.x

🚀 Getting Started

1. Prerequisites

2. Installation

git clone https://github.com/your-repo/biotrial-analytics.git
cd biotrial-analytics
npm install

3. Development

npm run dev

The app will be available at http://localhost:3000.


👤 Developer

Vivek Das | Expert Functional Prototype Development

Important

Disclaimer: This application is a functional prototype. All data and power estimates are simulated and should not be used for actual clinical decision-making without independent statistical validation.

About

This is a protoype biomarker analysis tool commonly used in trial developed using Gemini 3 and deployed using Google AI Studio

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors