Shuangsang Fang1
1BGI-Research, Belgrade, Serbia
Abstract
The large field-of-view and high-resolution spatial transcriptomics technology can reveal and answer scientific questions that cannot be discovered or elucidated by low-resolution spatial transcriptomics. Obtaining expression profiles at the single-cell level from high-resolution spatial transcriptomics requires sophisticated data processing and interpretation strategies, including extensive image data processing, transcriptome data processing, integration analysis. At the same time, the introduction of spatial information helps with the annotation of single cells at the tissue level and the study of tissue structure and function, while cell clustering and cell annotation are important foundations for subsequent in-depth analysis. Cell annotation can be divided into clustering and re-annotation based on marker genes and end-to-end cell annotation based on reference datasets. The choice between the two depends on whether markers are easier to obtain or whether reference datasets with consistent data backgrounds are easier to obtain. The algorithm team at BGI Research Institute has conducted extensive algorithm research and development on data interpretation strategies, cell clustering algorithms, and cell annotation algorithms for large field-of-view and high-resolution spatial transcriptomics technology, with the aim of providing comprehensive, efficient, and highly reliable data analysis algorithms, tools and platform support.
