The ASW in SDMBench is calculated as follows:
def compute_ASW(adata,pred_key,spatial_key='spatial'):
d = squareform(pdist(adata.obsm[spatial_key]))
return silhouette_score(X=d,labels=adata.obs[pred_key],metric='precomputed')
It appears to be calculated by inputting the distances between spatial positions into sklearn.metrics.silhouette_score. However, I am unable to reproduce the results for DLPFC12, including those from the STAGATE model. The results obtained using this method mostly fall within the range of -0.1 to 0.1, which significantly deviates from the results reported in the article. This discrepancy is similar to the results in the tutorial.
https://github.com/zhaofangyuan98/SDMBench/blob/main/Tutorial/tutorial.ipynb
The ASW in SDMBench is calculated as follows: def compute_ASW(adata,pred_key,spatial_key='spatial'): d = squareform(pdist(adata.obsm[spatial_key])) return silhouette_score(X=d,labels=adata.obs[pred_key],metric='precomputed') It appears to be calculated by inputting the distances between spatial positions into sklearn.metrics.silhouette_score. However, I am unable to reproduce the results for DLPFC12, including those from the STAGATE model. The results obtained using this method mostly fall within the range of -0.1 to 0.1, which significantly deviates from the results reported in the article. This discrepancy is similar to the results in the tutorial. https://github.com/zhaofangyuan98/SDMBench/blob/main/Tutorial/tutorial.ipynb