nnSVG for the scalable identification of spatially variable genes using nearest-neighbor Gaussian processes

Abstract

Feature selection to identify spatially variable genes is a key step during analyses of spatially-resolved transcriptomics data. Here, we propose nnSVG, a scalable approach to identify spatially variable genes based on nearest-neighbor Gaussian processes. Our method (i) identifies genes that vary in expression continuously across the entire tissue or within a priori defined spatial domains, (ii) uses gene-specific estimates of length scale parameters within the Gaussian process models, and (iii) scales linearly with the number of spatial locations. We demonstrate the performance of our method using experimental data from several technological platforms and simulations. A software implementation is available at https://bioconductor.org/packages/nnSVG.

Probabilistic embedding, clustering, and alignment for integrating spatial transcriptomics data with PRECAST

“…Supplementary Fig. S21 confirms the robustness of PRECAST by using top genes identified by different methods, such as SPARK 40 , SPARK-X 41 , SpatialDE 42 , and nnSVG 43 , as input.…”

The gene expression landscape of the human locus coeruleus revealed by single-nucleus and spatially-resolved transcriptomics

“…Alternatively, using standard significance thresholds of FDR < 0.05 and expression FC > 2, we identified a total of 437 statistically significant genes (Supplementary Figure 7B and Supplementary Table 2). As a second approach to identify genes associated with LC-NE neurons in an unsupervised manner, we applied a method to identify spatially variable genes (SVGs), nnSVG [38]. This method ranks genes in terms of the strength in the spatial correlation in their expression patterns across the tissue areas.…”

“…We applied nnSVG [38], a method to identify spatially variable genes (SVGs), in the Visium SRT samples. We ran nnSVG within each contiguous tissue area containing a manually annotated LC region (13 tissue areas in the N =8 Visium samples) and calculated an overall ranking of top SVGs by averaging the ranks per gene from each tissue area.…”

Disparities in spatially variable gene calling highlight the need for benchmarking spatial transcriptomics methods

“…SVGs are defined as having expression levels across a tissue that covary in a location specific manner (12,14). Published methods for SVG identification employ different mathematical models aiming to capture biological truth (12,13,(15)(16)(17)(18)(19)(20)(21)(22)(23). Benchmarking analysis tools is needed to ensure the reliability of processed data matches or supersedes that of similar technologies such as single-cell (sc) RNA-Seq (24,25).…”