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.
- 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.
- 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.
- 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:
$G_{lon} = 1 + (T-1)\rho$ .
-
Variance Decomposition: Subjects (
- Tissue Dynamics: Interactive Canvas-based simulation of tissue remodeling and immune infiltration.
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."
- 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
- Node.js (v18.0+)
- NPM or Yarn
git clone https://github.com/your-repo/biotrial-analytics.git
cd biotrial-analytics
npm installnpm run devThe app will be available at http://localhost:3000.
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.