Table 1.
Methodological improvements and new features in RCA2 compared to the prototype release
| Rca | RCA2 | |
|---|---|---|
| Reference panels | 1 | 11 |
| Species supported | Human | Human, mouse |
| Data import | Count matrix | 10x file import, Count matrix |
| Custom panel usage | Yes | Yes |
| Custom panel generation | Not supported | Automated |
| Clustering | Hierarchical (basic) | Hierarchical (divide-and-conquer), graph-based |
| Data structures | Standard | Sparse data structures |
| Clustering in | full projection space | full or PCA reduced projection space |
| Cell type/cluster specific QC | No | Yes |
| Parallelized correlation computation | No | Yes, if applicable |
| UMAP visualizations | 2D | 2D,3D |
| Cell type annotation | Not supported | Yes |
| Easy integration in existing workflows | No | Yes |
| DEG calling | No | 1 vs all and pairwise |
| Enrichment analysis | None | KEGG and GO |
| Number of UNIT tests | 0 unit tests | 90 unit tests |
| Tutorials and documentation | Limited documentation | Extensive manual and R-vignette |
| Applicable to datasets >20 000 cells | No | Yes |