Experimental-data-based in-silico data generation platform to improve the accuracy and reliability of single-cell and spatial omics data analysis

Project Summary This project aims to develop a suite of advanced yet practical statistical tools with user-friendly interfaces to enhance the reliability and power of single-cell and spatial omics data analysis through experimental-data- based in silico data generation. Aim 1 focuses on developing statistical methods to generate in silico data that serve as negative controls and pseudo-replicates of experimental data. These digital alternatives will help uncover potential biases and variability in analysis results, which have become more common given the increasing complexity of single-cell and spatial omics data analysis. In silico negative controls and pseudo- replicates will enable sanity checks, bias correction, and variability analysis, addressing challenges such as double dipping, small sample sizes, and data sparsity. Aim 2 involves creating a power analysis suite leveraging experimental-data-based in silico data generation, covering multi-condition comparisons, temporal data analysis, and population-scale molecular quantitative trait loci analysis, with the goal of assisting experimental design considering the high cost of single-cell and spatial omics technologies. Aim 3 will develop interactive, modularized software packages with a website interface for the single-cell and spatial omics community to perform experimental-data-based in silico data generation. The software will integrate with state- of-the-art pipelines like R's Seurat and Python's Scanpy, enabling researchers to easily generate in silico data from experimental data and enhance the reproducibility of common analysis tasks in single-cell and spatial omics studies. Overall, this project will provide a new angle to extend the capabilities of computational genomic research, fostering more accurate and reproducible data-driven discoveries.