Updating alteration data w.r.t. updated label

[ugurdogrusoz]

Looks like we're not paying attention to updated labels when it comes to alteration data. For instance, if you load some alteration data when a node's label is incorrect and then change the node label to the correct value, you expect the alteration data to be shown for the new label. Similarly, if you add a new node to the pathway after loading experiment data, you expect the data to be shown as well. Neither seems to happen right now!

[ugurdogrusoz]

One option here is to automatically update alteration data upon changes (e.g. addition of a new gene or label change of a gene) or we could provide a new menu item to "re-load/update" alteration data. Let's discuss.

[leonarddrv]

Addition of this functionality is not straightforward for both approaches (updated labels case and menu option) because the previous queries to retrieve genomic data are not saved. Also the structure of collaborative mode data need to be changed accordingly.

[ugurdogrusoz]

Thx, lets get back to this later then

[ugurdogrusoz]

@msalihaltun What do you think about this issue? If label changes cannot be directly caught and experiment data updated accordingly, can we perhaps add an Update alteration data menu item or something?

[msalihaltun]

The earlier comments are correct. We don't remember specific params about currently loaded queries (which data types we loaded from which study). This would not be a straightforward change.

[ugurdogrusoz]

How about an Update menu item / tool in the toolbar? That is, updating only when the user explicitly applies this operation? Is this easily doable?

[msalihaltun]

No, it wouldn't be. The same problem would occur.

Essentially, there are two issues. First we would need to remember the currently loaded study parameters (study ID and which boxes were ticked, e.g Mutation etc.). This requires some changes in the design, including collaborative data model. This should be doable, though I'm not sure how much needs to change. There are some aspects of how we keep the data (grouping and such) that I'm not very familiar with.

Second, we probably either need to remove all data and make a new query with the current list of nodes or just for the nodes that the change was made for (which I am not sure is doable). Removing all data from the DB would be a first step, but we should also make sure there are no race conditions (basically we should make sure the data is deleted from the DB before making any queries again) or other problems(?) here. There could be a bunch of things to consider.

This is why I'm hesitant to say it is easy.