StemnessIndex
First, stem cells were extracted to build a model by one-class logistic regression machine learning algorithm. Then, regression coefficients of transcriptome of stemness were obtained. Then, a given sample can be assessed based on the reference model. And a CSscore represents the correlation between the given sample and reference. The cell stemness scores were normalized within 0 to 1.
File format of expression profile:
I.The matrix must be genes (rows) by cells (columns). The row names should be the cell IDs and the column names should be the gene names.
II.The data must be delimited by tabs.
III.Uploaded data CAN NOT contain negative values. The normalization of TPM is acceptable.
Column names of sample information file should be as same as example is.
First, stem cell samples and normal samples from different tissues are used to develop the stemness index based on REO (relative expression orders), and gene pairs with certain expression orders (Gi > Gj or Gi < Gj) are retained in at least 99% of stem cell samples. REO (Gi < Gj or Gi > Gj) with orders inversion in at least 99% of normal adult tissue samples is taken as reference REO. Finally, the CSscore is the ratio of gene pairs with the same expression orders as the reference REOs to the total number of reference REOs. The cell stemness scores were normalized within 0 to 1.
File format of expression profile:
I.The matrix must be genes (rows) by cells (columns). The row names should be the cell IDs and the column names should be the gene names.
II.The data must be delimited by tabs.
III.Uploaded data CAN NOT contain negative values. The normalization of TPM, FPKM, RPKM or count are acceptable.
Column names of sample information file should be as same as example is.