Table 2.
seRNA-seq and spatial transcriptomic integration strategies.
| Integration strategies | Methods | Advantages | Disadvantages | Re |
|---|---|---|---|---|
| Deconvolution | SPOTlight, CellPhoneDB |
High accuracy | Does not incorporate capture location information when modeling spatial decomposition | (65) |
| Deconvolution | Cottrazm | Provide spatial quantitative information on cell composition | Highly dependent on the quality and completeness of reference data | (66) |
| Deconvolution | CARD | More precise | High computational complexity | (67) |
| Deconvolution | cell2location | Absolute quantification, not relative proportion | It has a high computational complexity and is extremely time-consuming | (68) |
| Deconvolution | cell2location | This provides strong and quantifiable evidence of spatial composition. | The technical deviation that cannot be completely avoided and lack the standard verification | (69) |
| Deconvolution | cell2location | Absolute quantification | Highly dependent on the quality and matching degree of reference data | (70) |
| Deconvolution | RCTD | Greatly enhance the detection sensitivity and deconvolution accuracy for target cell types, especially rare subtypes | RCTD will force the entire expression signal of each bin to be attributed to a combination of fibroblast subtypes | (71) |
| Deconvolution | SPOTlight | Higher resolution, capable of revealing cellular interactions | the high spatial heterogeneity among samples | (72) |
| Deconvolution | SPOTlight MIA |
No external reference data is required | It may confuse cell types and states | (73) |
| Deconvolution | cell2location | It can handle the inherent over-dispersion and technical noise in single-cell and spatial data very well, and the results are more robust and reliable | Biological verification is still required | (74) |
| Deconvolution | CARD MISTy |
The functions complement each other perfectly, forming an analytical closed loop | The accuracy of MISTy analysis is highly dependent on the accuracy of RCTD deconvolution | (75) |
| Deconvolution | CARD | Hierarchical annotation strategy improves accuracy | The recognition ability is limited and it is unable to parse new cell states | (76) |
| Deconvolution | SPOTlight, CellTrek |
Through multi-level and multi-angle verification, the conclusion is extremely robust | The analysis process is extremely complex and requires extremely high professional knowledge | (77) |
| Mapping | Tangram | Compatible with capture and image-based ST data | Gene expression can be less accurately predicted from histology images if the cells cannot be segmented | (59) |
| Mapping | CellTrek | Capture the complex nonlinear relationship between gene expression and spatial position | The spatial position of cells is predicted by the model rather than directly measured through experiments | (78) |
| Mapping | CellTrek | Realize spatial mapping at the single-cell level | high requirement for data matching degree | (79) |
| Mapping | CellTrek | true single-cell resolution spatial mapping | It is required that the scRNA-seq data and ST data must be derived from highly similar biological backgrounds | (80) |
| Spatially informed ligand–receptor analysis | SpaOTsc | The majority of cells can be mapped accurately using a small number of genes. | gnores the possible time delay associated with cell-to-cell communication | (81) |