ExtractHighCorFeatures {CountClust} | R Documentation |
This function compares grades of membership profile for each cluster in GoM model fit with the data expression profile to identify genes that are mostly strongly associated with each topic.
ExtractHighCorFeatures(omega, data, num_genes = 100)
omega |
\boldsymbol{omega} matrix, the relative grades of memberships from the GoM model fitting (a N x K matrix where N is number of samples, K number of topics). |
data |
G x N matrix of the expression profile of genes across samples, where G is the number of features and N number of samples |
num_genes |
The number of top associated genes with each cluster. Defaults to 100 |
A list containing two items - a K x num_genes matrix of the top strongly associated/correlated indices/features for K clusters, and another K x num_genes matrix of the absolute values of the correlations.