1) Query Results

The data base can be queried directly based on a combination of Cancer, Gene, and/or Drug combination. For selection of Cancer and Drug entries, one will limit the other based on the current TCGA data set. Not all Cancers will show a corresponding entry for all Drugs and vice versa. All generated interaction lists are sorted by raw p-value. For more detailed information on our methodology please see our citations.

For a specific Cancer-Drug-Gene interaction of interest, select each from the drop down menus located on the Interaction page and then click submit query. This will generate a page containing KM survival plots, expression data (if available), and Bonferroni corrected p-values. If no specifications are entered, a list of sorted interactions across all Cancer-Drug pairs will be generated. If only a Cancer is selected, then a ranked list of all interactions in this Cancer is produced. Similar lists are produced for only selecting a Drug or Gene of interest, where only interactions involving the selected parameter are listed. If a specific Cancer-Drug pair is selected all genes in that analysis are displayed, with a Bonferroni correction cutoff shown by highlighting the genes found to not be significant in red. If a Cancer-Gene or Drug-Gene pairing is selected a list of all interactions by Drug and Cancer respectively are shown.

Selction dropdown lists for cancer and drugs are dynamically updated upon the user's selections. When a user selects a cancer, the drug list is updated in such that only durgs with 30 or more patients are shown. When the user selects All in the cancer dropdown list, the drug list is restored to show all drugs in the data. Similarly, when a user selects a drug, only cancers with 30 or more patients exposed to the drug are shown in the cancer dropdown list. When the user selects All in the drug dropdown list, the cancer list is restored to show all cancers. This dynamic update is implemented based on the number of patients exposed to each drug in each cancer type, and a summary table of the patient counts is available here: DrugExposureSummary_All.csv

Entry	Cancer	Drug	Gene	Output
Input	All	All	All	Return P-value ranked list of all significant results.
	x	All	All	All available Drug-Gene pairs found in this Cancer.
	All	x	All	All available Cancer-Gene pairs found in this Cancer.
	All	All	x	All available Cancer-Drug pairs found in this Gene.
	x	x	All	All Genes in this Cancer-Drug pair.
	x	All	x	All-Cancer-Drug pairings with this Gene, limited by number of patients with drug exposure
	All	x	x	All Cancers with this Drug-Gene pairing.
	x	x	x	Drug stratified survival analysis of this Cancer-Drug-Gene interaction.

2) Select Cancer Drug pair

We have provided a list of all 77 Cancer-Drug pairs that have atleat a minimum number of patients for survival analysis. Clicking the blue-highlighted number of significant interactions in the provided table takes the user to a sorted list of those interactions. Each number means the number of important Genes of that particular Cancer-Drug pair. From there, clicking the link option for a given interaction will take the user to the survival and expression plots.

3) Heat Map

Click the Heat-Map link will take the user to an abbreviated heat map of the number of patients per Cancer-Drug pairing. This heat map includes all pairings that have at least the minimum number of patients needed for analysis, with many that fall below this threshold removed. Clicking the blue-highlighted number of patients with take the user to the Cancer-Drug query as described above.