STmulator

Simulation and interpolation of spatial transcriptomics from parameter cloud

STmulator: Simulation and interpolation of spatial transcriptomics from parameter cloud

Status: Active project, ready for RECOMB 2026 conference submission

Role: First author, Independent Research Project

Institution: Prof. Maizie Zhou Lab, BME & CS Department, Vanderbilt University

Duration: August 2024 – Present (Remote and on-site collaboration)

Project Overview

STmulator is an innovative statistical modeling framework designed to simulate and interpolate spatial transcriptomics (ST) data using advanced mathematical techniques. This project addresses the critical need for realistic synthetic spatial transcriptomics data for method development, validation, and hypothesis testing in the rapidly evolving field of spatial genomics.

Technical Innovation

The core methodology involves:

Statistical Modeling:

  • Vine-Copula Modeling: Advanced statistical modeling of parameter space relationships
  • Single Slice Simulation: Realistic generation of individual spatial transcriptomics slices
  • 3D Region Simulation: Extension to multiple slice and three-dimensional tissue region simulation

Interpolation Framework:

  • Optimal Transport Theory: Mathematical foundation for interpolation between spatial states
  • Geodesic Path Computation: Principled interpolation along optimal transport-driven geodesic paths
  • Parameter Cloud Navigation: Efficient sampling and navigation through high-dimensional parameter spaces

Complete Research Cycle

This project represents a comprehensive research experience including:

  • Literature Review: Extensive survey of spatial transcriptomics simulation methods
  • Idea Exploration: Creative hypothesis generation and initial concept development
  • Failure Analysis: Learning from initial approaches that didn’t meet expectations
  • Method Validation: Rigorous testing and validation of novel approaches
  • Model Construction: Implementation of robust, scalable computational framework
  • Benchmark Development: Creating comprehensive evaluation metrics and comparison studies
  • Application Studies: Demonstrating utility across diverse biological scenarios
  • Manuscript Preparation: Scientific writing and presentation of results

Professional Development

The project has provided extensive training in:

  • Public Presentation: Regular presentation of progress at different project milestones
  • Scientific Communication: Development of clear, compelling research narratives
  • Independent Research: Self-directed project management and problem-solving
  • International Collaboration: Remote and on-site work with US-based research team

Expected Impact

STmulator will provide the research community with:

  • Realistic ST Data Generation: High-fidelity synthetic spatial transcriptomics datasets
  • Method Benchmarking: Standardized evaluation framework for ST analysis methods
  • Hypothesis Testing: Platform for testing spatial biology hypotheses in controlled settings
  • Educational Resource: Tool for teaching spatial transcriptomics concepts and methods

Conference Submission

The project is currently being prepared for submission to RECOMB 2026, one of the premier conferences in computational biology, demonstrating the high quality and significance of this research.

This project showcases independent research capability and innovative thinking in computational biology.

References