in your article,you saied use rPCA to intergrate all sample's tumor spot,then use like snn method clustered into 6 distinct cluster and use findallmarkers function to find cluster's marker gene.
but in your Rscript Module_Analysis:Shared Transcriptional Programs ,you used a totally different way to find these 6 clusters,first run clustering for each tumor sample ,and find cluster's marker gene for each sample ,then you filtered these marker gene by a (fold change>2 & top50 frequency across all sample ) way.then you use these filtered gene to subset all_sample_merged spot-gene matrix,and use this subset matrix to cor,then hclust this cor matrix ,and you come to 6 clusters
my question is why your article describe is totally different with your R script code.
in your article,you saied use rPCA to intergrate all sample's tumor spot,then use like snn method clustered into 6 distinct cluster and use findallmarkers function to find cluster's marker gene.
but in your Rscript Module_Analysis:Shared Transcriptional Programs ,you used a totally different way to find these 6 clusters,first run clustering for each tumor sample ,and find cluster's marker gene for each sample ,then you filtered these marker gene by a (fold change>2 & top50 frequency across all sample ) way.then you use these filtered gene to subset all_sample_merged spot-gene matrix,and use this subset matrix to cor,then hclust this cor matrix ,and you come to 6 clusters
my question is why your article describe is totally different with your R script code.