SpatialTranscriptomicsTutorials

Tutorials of spatial transcriptomics data loading, including 10X Visium and Old ST. 目前，大多数的公开空转数据集，在这两个空间转录组数据集收集网站上可以找到**(.h5ad)**为文件后缀：

1. STOmics DB：中国国家基因库数据库
2. SODB：复旦大学与腾讯智能医学院的联合项目。

Spatial Domain Identification Methods可大致分为三类：

Current methods for identifying tissue structures for spatial transcriptomics data include three categories:

1. Clustering using gene expression profiles alone, the representative method is Louvain (implemented by SCANPY and Seurat);
2. Clustering using both gene expression profiles and spatial location, the representative method is BayesSpace;
3. Clustering using gene expression profiles, spatial location and histological image, the representative method is SpaGCN.

下面是各个分类中的详细代表性算法：

1. Non-spatial:

   (1) SCANPY (Genome Biology 2018.02)

   (2) Seurat (Nature Biotechnology 2015.04 (V1), Nature Biotechnology 2018.04 (V2), Cell 2019.06 (V3), Cell 2021.05 (V4), Nature Biotechnology2023.05 (V5))

2. Spatial but no morphology:

   (1) Giotto (Genome Biology 2021.03)

   (2) BayesSpace (Nature Biotechnology 2021.06)

   (3) DR-SC (NAR 2022.07)

   (4) STAGATE (Nature Communication 2022.04)

   (5) BASS (Genome Biology 2022.08)

   (6) BANKSY (bioRxiv 2022.04)

   (7) SpatialPCA (Nature Communication 2022.11)

   (8) SEDR (bioRxiv, 2021.06)

   (9) SOTIP (Nature Communication 2022.11)

   (10) GraphST (Nature Communication 2023.03)

3. Spatial and morphology:

   (1) stLearn (bioRxiv 2020.05)

   (2) SpaGCN (Nature Methods 2021.10)

   (3) DeepST (NAR 2022.10)

   (4) ConST (bioRxiv 2022.01)

此外，还有比较奇葩的，比如用了形态学和转录信息，但是没有用到空间信息的：

1. ConGI (BIB 2023.02)
2. MUSE (Biotechnology 2022.05)