Annotation Modules
This section corresponds to different anno_algorithm branches under option=annotation and provides several annotation strategies, including manual annotation, reannotation, cell2location, and RCTD.
Please make sure that you have already completed preprocess, clustering, and annotation_help, and that the annotation support files are available.
For a detailed explanation of the shared annotation configuration template, see annotation.yaml Reference.
We recommend choosing the annotation method according to your dataset characteristics and analysis goals. If you are familiar with the expected cell types and your dataset is relatively small, manual annotation is often sufficient. If your dataset is larger, or you require a more systematic annotation strategy, cell2location or RCTD may be more appropriate.
In general, cell2location is better suited to lower-resolution spatial transcriptomics data, whereas RCTD is more appropriate for higher-resolution datasets.
If matched single-cell data are available, you can use your own sequencing data as the annotation reference in a multimodal analysis setting. If you do not have matched single-cell data, you may instead use a public single-cell reference dataset.In that case, make sure the public dataset is reliable and well annotated, and provide the correct cell type annotation column name together with the path to the reference annotation results.