 NNSVG is a novel algorithm for identifying spatially variable genes from spatially resolved transcriptomic data. It uses nearest neighbor Gaussian processes to estimate the variation of gene expression across space and can be used to identify genes whose expression varies continuously over the entire tissue or within predefined regions. The algorithm is scalable and requires only a few minutes to run on large datasets. Additionally, it is able to accurately estimate the length-scale parameter of each gene, which allows for more accurate estimation of the variance of gene expression. Finally, the algorithm is computationally efficient and scales linearly with the number of spatial locations. This article was authored by Lukas M. Weber, Akajo-Otisaha, Abhayrup Dutta, and others.