WashU Pan-Cancer miRNome Atlas


Search by miRNA
  Tumor Development
  Tumor Stage and Grade
  • Survival Outcome
  • Expression Profile
Search by Cancer Type
  Tumor Development
  Tumor Stage and Grade
  Survival Outcome
  • Expression Profile
Search for miRNA-Target Correlation
  Search by miRNA
  Search by Gene
Survival Signature Analysis
  Pre-calculated signatures
  Custom signatures
Clustering Analysis
  Hierarchical clustering
  K-means clustering
Cluster cancer types by miRNA expression

Select at least 3 cancer types
Select all cancers
Use all miRNAs or include a list of miRNAs for clustering.
Use all miRs within filter range in creating clusters
Please submit miRNAs using 3p/5p designations. Separate with spaces, semicolons, commas or newlines.
Example submission
Additional cluster options
Number of iterations for k-means clustering (max 25) 
Input the number of clusters of patients to show in tabular format  
Filter type:       Minimum:      Maximum: 

Use k-means clustering to group cancer types by miRNA expression.

K-means clustering separates samples by relative distance between features, which here are miRNA expression levels.
Provide a list of miRNAs to use in clustering in the text box, or check the box to use all available miRNAs.

A consensus of clustering runs can be selected by increasing the number of iterations. The number of clusters must be specified.

miRNAs can be excluded on the basis of mean expression or standard deviation of expression throughout the dataset. Use the minimum/maximum values to set the filter limits.