About
SPARK is a statistical method for identifying spatial expression patterns. SPARK directly models raw count data generated from various spatial resolved transcriptomic techniques. With a new efficient penalized quasi-likelihood based algorithm, SPARK is scalable to data sets with tens of thousands of genes measured on tens of thousands of samples. Importantly, SPARK relies on newly developed statistical formulas for hypothesis testing, producing well-calibrated p-values and yielding high statistical power. We illustrate the benefits of SPARK through extensive simulations and in-depth analysis of four published spatially resolved transcriptomic data sets. In the real data applications, SPARK is to 10-fold more powerful than existing approaches and identifies new genes that reveal the importance of neuronal migration in the formation of the olfactory system as well as the importance of immune system and cytoskeleton stability in tumor progression and metastasis.
SPARK-X is a a scalable non-parametric method that can effectively and rapidly identify genes with spatial expression patterns in large spatial transcriptomic studies. SPARK-X builds upon a robust covariance test framework, producing effective type I error control and high statistical power in simulations. In the real data applications, SPARK-X is orders of magnitude faster, requires orders of magnitude less memory, produces calibrated type I error control, and identifies many new genes not identifiable by existing approaches. SPARK-X represents an effective complement of existing spatial expression detection methods towards analyzing large spatial transcriptomic data that are being collected today.
Cite SPARK
Shiquan Sun*, Jiaqiang Zhu* and Xiang Zhou#. Statistical analysis of spatial expression pattern for spatially resolved transcriptomic studies, 2020, Nature Methods, in press.
Jiaqiang Zhu, Shiquan Sun and Xiang Zhou#. Scalable and robust non-parametric detection of spatial expression patterns for large spatial transcriptomic studies., 2020