The following is a summary of “Dimension-agnostic and granularity-based spatially variable gene identification using BSP,” published in the November 2023 issue of Nephrology by Wang et al.

Identifying spatially variable genes (SVGs) is pivotal in establishing connections between molecular cellular functions and tissue phenotypes. Spatially resolved transcriptomics allows for the capture of gene expression at the cellular level alongside corresponding spatial coordinates in two or three dimensions, offering a promising avenue to infer SVGs effectively. However, existing computational methods often struggle to deliver reliable outcomes, particularly in handling three-dimensional spatial transcriptomic data.

This study presented BSP (big-small patch), a non-parametric model that compares gene expression patterns at two distinct spatial granularities. BSP aimed to swiftly and robustly identify SVGs from two or three-dimensional spatial transcriptomics data. The chosen method has undergone rigorous testing in simulations, showcasing superior accuracy, robustness, and remarkable efficiency.

The effectiveness of BSP has been further validated through substantial biological discoveries across diverse domains, including cancer, neural science, rheumatoid arthritis, and kidney studies. BSP has proven its efficacy in analyzing various spatial transcriptomics technologies, reinforcing its versatility and reliability.

By introducing BSP, the researchers offered a potent tool for seeking to uncover SVGs from complex spatial transcriptomic datasets. Its ability to handle multi-dimensional data swiftly and accurately, coupled with its validated performance across different biological contexts, underscores BSP as a promising approach for identifying spatially variable genes with significant implications for understanding cellular behavior within tissues.

This innovative BSP (big-small patch) model significantly advanced spatial transcriptomics analysis. Its robustness in handling multi-dimensional data and demonstrated accuracy across diverse biological studies positions BSP as a valuable tool for researchers investigating spatially variable genes, promising deeper insights into complex tissue dynamics and cellular behavior.

Source: nature.com/articles/s41467-023-43256-5