Instead of having a pulldown with datasets, we can have a single 'datasets' page that describes for each dataset:

1. publication of origin
2. data modalities available, what was done to data

• for each data type, how it was transformed, e.g drug data, counts to TPM, etc.

1. total number of samples x drugs

• rnaseq + drug
• rnaseq+mutation+drug
• rnaseq+copynumber+drug
• mutation+copynumber+drug

1. Drug curve metrics collected:

• for each metric, how many drugs do we have? (this is related to Train_test_validate function does not work for nci60 dataset #322)

1. recommended metric?