baliga-lab / GGBWeb

Integrative, web-based genome browser based on the Gaggle Genome Browser
MIT License
0 stars 0 forks source link

#8 - ChromeGoose integration [to implement] #39

Open dsalvanha opened 9 years ago

dsalvanha commented 9 years ago

Ok. The scenario we have now demands us to have some updates on already implemented features:

on Module: Module itself has a configuration file that sets all lanes (and its tracks) to be rendered. _IMPORTANT_\ : There can only be 1 Module configuration set at a time !!!!!!! Whenever user wants to open a gene in a different module we NEED to ask WHETHER or not the user wants to change the Configuration.

two scenarios: If the user chooses to change the configuration ALL WINDOWS should change to the new module configuration. (lanes with its tracks) If the user chooses NOT TO change the configuration, the new gene will open with current configuration

on init of each GGBWindow We should have a configuration argument now to GGBWindow

if configuration is empty (most of the time that user does not use chromeGoose) - we open it with default or last state. if configuration is set to a root module item --> we open GGBWindow with all available tracks.

A first thought on this configuration file is:

configuration = { module_item_name:{ Heatmap:{exp1, exp2, exp3}, QuantitativeSegment:{exp4, exp5} } }

With this, we should open GGBWindow with 3 lanes (default, heatmap, quantitativeSegment) and each experiment populated corretly inside each lane

Only a root module item has a configuration and ALL it's sub-module elements follows it

This feature will be REALLY powerfull.

njiang commented 9 years ago

Go to a TB module page such as: http://networks.systemsbiology.net/mtb/biclusters/bicluster_0001

Chrome goose will recognize the data on the page. Select "MTB filtering" from the Data Processing/Filter menu. Then in the dialog box, select the bicluster genes as input and click "run". The data will be passed to the opencpu server. Once it is finished, results will be passed to the "handleGooseData" function. We can then add the data to the module page and render them.

dsalvanha commented 9 years ago

Cannot get results from openCPU using the test location group for e. coli (attached)

Error I get: image

dsalvanha commented 9 years ago

Look what I get on Boss Clipboard:

image

Why r we broadcasting only the element name instead of all other informations ?

njiang commented 9 years ago

You should use TB genes instead of Ecoli genes. And you deployed opencpu on your local server? Please install the latest Chrome goose and it should point to the server in the Amazon cloud.

njiang commented 9 years ago

Let's clarify the "submodule" concept. So do you mean each module has a list of genes (same as now), and each gene has a list of genes (or tracks) corresponding to it? If it's a track, how do we get the track information? In the case of TB integration, we receive genes together with their overlapped genes in biclusters etc. We need a better way to display the data.

If user clicks to open a gene in a new window, we should render the "root" gene as well as all its associated genes using the configurations? When and how do we generate the configurations?