tiagochst
commented
2 years ago Hi, there is a function in TCGAbiolinks called TCGAanalyze_Stemness https://rdrr.io/bioc/TCGAbiolinks/man/TCGAanalyze_Stemness.html
But here is the code for stemless score (you can adapt it) and data (I added it in https://github.com/BioinformaticsFMRP/PanCanStem_Web/tree/master/Stemsig):
signature <- readr::read_tsv(
"https://raw.githubusercontent.com/BioinformaticsFMRP/PanCanStem_Web/master/Stemsig/SC-pcbc-stemsig.tsv",
col_names = F
)
signature.weight.vector <- signature$X2
names(signature.weight.vector) <- signature$X1
# Just an example with correlation 1 and -1
gene.expression.matrix <- matrix(signature$X2)
rownames(gene.expression.matrix) <- signature$X1
gene.expression.matrix <- cbind(gene.expression.matrix,gene.expression.matrix * -1)
calculate_score <- function(signature.weight.vector, gene.expression.matrix){
# Keep only common genes
common.genes <- intersect(names(signature.weight.vector), rownames(gene.expression.matrix))
gene.expression.matrix <- gene.expression.matrix[common.genes, ,drop = FALSE]
signature.weight.vector <- signature.weight.vector[common.genes]
score <- apply(gene.expression.matrix, 2, function(sample) {
cor(sample, signature.weight.vector, method = "sp", use = "complete.obs")
})
print(paste0("Min score: ",min(score)))
print(paste0("Max score: ",max(score)))
# Scale the scores to be between 0 and 1
print(paste0("Normalized scores to be between 0 and 1"))
score <- score - min(score)
score.normalized <- score/max(score)
print(paste0("Min normalized score: ",min(score.normalized)))
print(paste0("Max normalized score: ",max(score.normalized)))
return(score.normalized)
}
calculate_score(signature.weight.vector,gene.expression.matrix)
Nuvolar
I can't find the 'how to document' where can I find ?
Thanks.