Abstract
Motivation Rigorous experimental design is essential for obtaining biologically meaningful findings and high-throughput spatial transcriptomics studies are not an exception. While some efforts have been made to improve the design of these studies, it is still significantly understudied yet how to optimize key experimental factors of these experiments. In the case of sequencing-based spatial transcriptomics studies, determining the minimum sequencing depth is an important experimental factor to decide.
Results To address this critical limitation, here we propose spaDesign, a statistical framework to improve the design of sequencing-based spatial transcriptomics experiments. spaDesign is a statistically rigorously designed framework that employs Poisson Gaussian process and Fisher-Gaussian kernel mixture. It can easily simulate a range of spatial transcriptomics data with various sequencing depths, effect sizes, and spatial patterns, which allows rigorous estimation of needed total sequencing depths to detect spatial domains based on spatial transcriptomics experiments. We demonstrated the utility and power of spaDesign using 10X Visium data of the human brain and the chicken heart.
Availability The R package and associated tutorial are freely available at https://github.com/JuanXie19/spaDesign.
Competing Interest Statement
The authors have declared no competing interest.
