| Data properties |
Spot-level |
Fraction zero |
No |
Kernel Density |
[0, infinity) then map to [0, 1] |
|
|
Library size |
No |
|
|
|
|
TMM |
No |
|
|
|
|
Effective library size |
No |
|
|
|
|
Scaled variance |
Yes, z-score standardization |
|
|
|
|
Scaled mean |
Yes, z-score standardization |
|
|
|
|
Library size vs Fraction zero |
No |
|
|
|
|
Sample Pearson correlation |
No |
|
|
|
Gene-level |
Fraction zero gene |
No |
|
|
|
|
Scaled variance |
Yes, z-score standardization |
|
|
|
|
Scaled mean |
Yes, z-score standardization |
|
|
|
|
Mean vs variance |
No |
|
|
|
|
Mean vs variance (scale) |
Yes, log-normal transformation |
|
|
|
|
Mean vs fraction zero |
No |
|
|
|
|
Gene Pearson correlation |
No |
|
|
|
Spatial-level |
Transition Matrix (TM) |
No |
|
|
|
|
Neighborhood enrichment matrix (NEM) |
No |
|
|
|
|
Centralized score matrix (CSM) |
No |
|
|
|
|
Cell type interaction |
No |
|
|
|
|
Moran’s I |
No |
|
|
|
|
L statistics |
No |
|
|
|
|
Nearest neighbor correlation |
No |
|
|
| Spatial downstream analysis |
Spatial clustering |
|
No |
Adjusted rand index (ARI) |
[-1, 1] |
|
|
|
No |
Normalized mutual information (NMI) |
[0, 1] |
|
Cell type deconvolution |
|
No |
Root mean square deviation (RMSE) |
[0, infinity) |
|
|
|
No |
Jensen-Shannon divergence (JSD) |
[0, 1] |
|
Spatially variable genes (SVG) identification |
|
No |
Precision |
[0, 1] |
|
|
|
No |
Recall |
[0, 1] |
|
Spatial cross-correlation |
|
No |
Cosine similarity |
[-1, 1] |
|
|
|
No |
Mental statistics |
[-1, 1] |
| Scalability |
200 x 200 |
|
No |
time |
[0, infinity) |
|
3k x 500 |
|
No |
memory |
[0, infinity) |
|
5k x 1k |
|
|
|
|