Nov 12 2011

Beautiful Data

It has been said that if one really wants to learn something, they should teach it. However we may want to expand that aphorism, to perhaps if one really wants to learn something, they should teach it, organize it or animate it.

The commonality between science and art is in trying to see profoundly – to develop strategies of seeing and showing. –Edward Tufte

One of the main goals of our Datapunk bioinformatics platform is the development of new and exciting information visualization (InfoViz) tools that can be used to develop new appreciations for the relationships between data. We are currently developing two new InfoViz platofrms that I think have great potential. But more than that, these tools feature stunning interfaces that go a long way towards again proving the fact that information, presented imaginatively, can not only yield amazing secrets, but can be stunningly beautiful as well.

Pathscrubber

The first InfoViz platform we developed is a full-bodied genomic network depictor called PathScrubber, which runs inside Datapunk. PathScrubber draws network graphs of gene-protein relationships. Each node in the network is click-able and links to a popup that provides information on that gene, through an API to OMIM (Online Mendellian Inheritance in Man). Perhaps more significantly, Datapunk is the first informatics tool that is harvesting scientific references detailing phytochemicals and dietary agents that have been reported in the literature to influence the expression of these gene-proteins.

PathScrubber.

Simply enter any gene-protein terms you wish to include in your network. Don’t worry about partial terms; PathScrubber will return a list of possible terms for you to consider and you can check which ones are appropriate in the next screen. After the program draws the network, you can zoom in or out with either the mouse or via a slider. Nodes containing genomic expression information on phytochemical or dietary agents are coded orange. PathScrubber has an extensive help page. PathScrubber is programmed in Perl, utilizing Léon Brocard’s GraphViz module to draw the graphs.

InfoViz Democratizers

This platform is designed to provide a venue to allow naturopathic physicians and researchers to animate their own data. One of the most exciting/important things we are developing is a set of guidelines for authors on how to structure their data so as to allow us to easily port it into a stunning visual display. Most JavaScript data is encoded in a format called JSON, which although paradoxically designed to be easily readable by humans when compared to other data structures, requires a heavy degree of ‘nesting’ of the data, which most non-programmers would find confusing and thus increase their proneness to data entry error. So we developed a simple language that only requires that the data be entered in a simple text file with a few codes. This is then parsed by Perl into JSON and piped to the page as HTML. Here are two examples:

Lectins: Classification and Taxonomy
Our second infoviz tool is a depiction of a the taxonomy and classification of known animal, plant and microbial lectins. Lectins are protein molecules that attach to sugars and modulate a variety of cell functions, including mitosis, agglutination, metastasis and infections. Most of the data for this infoviz is from my textbook, Fundamentals of Generative Medicine Clicking on a node should move the tree and center that node. This infoviz makes extensive use of JavaScript, especially the JavaScript InfoVis Toolkit. The tree-like structure opens and closes as one click on the various categories.

Lectin Classifications

Radio buttons on the top allow for the user to display the tree in different aspects and to chose between a ‘normal’ display, where the categories open up in a linear fashion of a ‘centering’ mode where the newly selected category moves to the center of the tree. Finally branches of the tree often contain hyperlinks to additional information.

Actions of Medicinal Plants
Our third infoviz tool is a depiction of a paper developed by Eric Yarnell, ND entitled ‘A Compendium Pharmacological Actions of Medicinal Plants and Their Constituents.’ Clicking on a node should move the tree and center that node. This InfoViz also makes extensive use of JavaScript, especially the JavaScript InfoVis Toolkit to display day in the form of a morphing hypertree. The centered node’s children are displayed in a relations list in the right column. The data set for this InfoViz is still in development.

Compendium of Medicinal Herbs

No responses yet

Comments are closed.