[x] The process of adding nodes is slow, we should figure out what part is slow (i.e. adding edges, getting current positions etc). There might also be a built-in delay in the addCyNodes function. -> Skipping edges cuts the time in half. Can't find anywhere in RCy3 code where there is a delay in addCyNodes function.
[x] Get all node positions at once, for two reasons: 1) it might be faster, 2) can use the information to cross-check if there are overlapping nodes during the placement of ptm nodes. -> Collecting all positions up front improves speed
[ ] Add a proper id mapping step using Ensembl protein ids for phospho data (Protein data has Ensembl gene IDs)
[ ] Alternate visualization: add pie chart to the right (?) of each node, with each part representing a ptm. More compact, could visualize more sites easily. Will need to use enhancedGraphics since Custom Graphics works with data columns. Create one new column with data points for relevant ptms. Question for Scooter: Can the list have different lengths for different nodes, aka different number of pie pieces?
[ ] Alternate visualization: Use rectangular "tags" on the right edge of the node to represent each ptm. Use 4 max. For 4 ptms, the tags would fill the full side.
[x] follow-up with Sarah on her real use case involving contrasting protein and phospho changes -> Sarah will ask about the possibility of using collaborator's unpublished data
[ ] Use the PROGENy pathway scores to narrow down which phospho sites to display. Sites with higher "activity score" are the ones that contribute to the pathway activity, and are thus more relevant to visualize.