EAGS: efficient and adaptive Gaussian smoothing applied to high-resolved spatial transcriptomics

The emergence of high-resolved spatial transcriptomic (ST) technology has led to novel approaches to investigating biological development, growth, and other complicated problems. With high-resolved ST data and highly sparse gene expression patterns, the most frequently employed approach for imputation has difficulty tackling the problem. We propose an efficient and adaptive gaussian smoothing (EAGS) approach for high-resolved ST. This is a nearest neighbor-based adaptive two-factor smoothing approach for genomic missing information repairing. The efficiency and effectiveness of the method on high-resolved ST data is demonstrated by validating it on ST slice data from mouse brain and olfactory bulb, compared with other competitive methods, EAGS has better partitioning accuracy and biological interpretation and possessing a significant advantage in computational power consumption.