BEASTsim: A Benchmarking and Analysis framework for Spatial Transcriptomics Simulations

BEASTsim (BEnchmarking and Analysis of Spatial Transcriptomics simulations) is an advanced benchmarking framework designed to assess the performance of various simulation techniques used in spatial transcriptomics. This tool provides standardized testing and evaluation metrics, enabling researchers to compare different spatial transcriptomics simulators and provides tools for analysis of such spatial data.

This framework is developed as part of the MOPITAS project, funded by the Novo Nordisk Foundation, which aims to develop experimental and computational methods to track single cells in space and time by integrating spatial transcriptomics and scRNA-seq.

Table of Contents

1. Features
2. Usage and Tutorials
3. Documentation
4. Acknowledgements
5. Contact

BEASTsim has two modules; benchmarking and analysis. The benchmarking module evaluates spatial transcriptomics simulation across data property estimation, biological signal preservation, and similarity-based metrics. The analysis pipeline allows for more in dept analysis of the simulated tissues such as cell type neighborhoods, spatially variable genes, tissue similarity, and more.

Usage and Tutorials

The tutorials covering the benchmarking and/or analysis of spatial transcriptomics simulation methods using BEASTsim, can be found here.

Please report any bugs via GitLab issues, and feel free to contact us if you have any questions regarding BEASTsim.

Documentation

Below is a table of contents to help you navigate the available documentation:

• Installation: Step-by-step instructions to install and set up BEASTsim.
• Parameters: Complete list of BEASTsim parameters with their type, default values, possible values, and detailed descriptions.
• Functions: Comprehensive documentation for all public BEASTsim functions and their usage.
• Style guide: Coding conventions and documentation guidelines for BEASTsim.

Acknowledgements

We thank all paper authors for their contributions: Tomás Bordoy García-Carpintero, Lucas A. D. T. Dyssel, Kristóf Péter, Nikolaj F. H. Hansen, Lena J. Straßer

We also would like to thank Chit Tong Lio, Merle Stahl, Markus List, and Richard Röttger for their valuable feedback on both paper and package.

Contact

• Tomás Bordoy García-Carpintero - Email: tobor@imada.sdu.dk
• Lucas Alexander Damberg Torp Dyssel - Email: ludys@imada.sdu.dk