VA_20100430

iPG2P Visual Analytics Minutes
April 30, 2010; 1 to 2 pm EDT

Present: Matt Vaughn, Adam Kubach, Eric Lyons, Lenny Heath, Steve Welch, Ruth Grene, Lecong Zhu, Bernice Rogowitz, Bjorn Usadel, Nick Provart, Tina Lee.

The meeting was convened at 1pm EDT

Item 1: Update on Working Group progress - Ruth and Bernice
Bernice shared her desktop, demonstrating some of the prototype work that she has been doing using ViVA and datasets. She looked at Bjorn's experimental data, trying to explore and ID potential interesting gene clusters under different conditions. First took look at p-values for each condition. Two replications, corrected for multiple testing. She tried to identify results that different from controls across all short conditions. She did the same thing for long experiment and found only 3 genes different, 2 of which were different from controls in green condition. Part two of the analysis was to identify genes differentially expressed under different conditions - looking for genes visually different from controls. She was able to paint different pictures for short and long experiments. 28 genes were highly expressed in long, cold condition. Of these, significantly different ones can be identified and she is now working towards Venn Diagram for coexpression.

Item 2: Discussion on progress - All
It is possible to to AGIs and annotations but it is a matter of doing the joins and then can share dataset and others. Bjorn added that there is downloadable data available from MapMan. This is interesting for biologists to see things visually. Provart stated that with Venn visualization, "Sungear" allows to do things with circles so Venn-like diagrams can be generated for any number of datasets.

Item 3: MapMan's capabilities - Bjoern
Bjoern shared his desktop, demoing MapMap's latest tools, biological design to parser/paths and visualization for any kind of species. The site has been moved to a new website where you can download applications and files. "Robin" ties into MapMap: load your data, graphically annotate experiments, when finished, MapMan automatically uploads and directly displays the data.

It can display time and condition series experimental data as it is agnostic relative to the data. Bjoern stated that the problem is in this tiny parser/paths. It is easy to visualize but gets messy as you get larger. Robin is a statistical analyzer. Bjoern showed that if you mouse over boxes, you can also access TAIR web services for the particular gene. It puts the abstract in the box for the particular article. It can also display a Venn diagram for n=4 or less.

Bernice asked if could use ViVA outputs and talk with MapMan. Yes, but it needs to be integrated and that is major work but it is possible.

Ruth asked if had a time course, how hard would it be to show dynamic changes. Bjoern said it would be need to do some statistical filtering. Bjorn's pathways do not have relationship to Keck directly.

Item 4: Assign Action Items