We recently attended an interdisciplinary visualization workshop that was all about creating a dialogue between scientists, technologists and designers. It was interesting to discuss the different ways in which these groups think about visualization and how they use it for different purposes. Very bluntly put, each group lacks something another group knows and cares deeply about, be it an understanding of colour [we met in the UK] or an understanding of statistics.
We had a deep discussion on interdisciplinarity, whether one person should feel comfortable and knowledgable enough doing work in all fields, or whether it’s really about people and communication skills that are needed to successfully collaborate with specialists in their own fields. Regarding scientists and visualization in particular, one point that stuck out was how scientists had to embrace programming as a tool for their experiments, even though they are not computer scientists. Will they have to embrace visualization techniques in the same way?
Of course, scientists already use visualizations in their work. Even though they may not be aesthetically the most pleasing, they do serve their purpose – or do they? Computer scientist Miriah Meyer and designer Bang Wong are doing research to uncover the unused potential of visualizations for scientists that can help them in understanding and validating their scientific data and also support them in developing new hypotheses and insights through visual tools.
Miriah strongly believes in the power of multidisciplinary teams instead of knowing how to do everything yourself. But still, she thinks that a basic understanding of visualization is very important for every undergraduate science student and should be taught, so they know it exists, what potential it has and that there are actually people they can ask to help them with doing research using visualizations.
Pathline is a tool that Miriah and Bang developed in close collaboration with geneticists applying a design process. They explored and designed before implementing anything and also incorporated user feedback into the tool. They didn’t use such a design process before and were very happy to discover, how it allowed them to move faster than if they had started the implementation part right away. What came out of this is a, in their own words, “visualization tool for comparative functional genomics that supports analysis of three types of biological data at once: functional data such as gene activity measurements; pathway data that presents a series of reactions within a cellular process; and phylogenetic data describing ancestral relationships between species.”
In addition to the above “curvemap”, the “linearized pathway” was developed. It has a very simple structure, always ranging from 0 to 1 from left to right. It can be enhanced by adding more detail to its structure.
This added detail may look more complex – but only if you don’t know what to look for. Our favourite [still trying to be British here] example for this is a topographic map: there’s an incredible amount of detail in there, but if you know that you’re looking for place A and B and the road between them, you’ll find them easily with a little bit of training, because you know how they are visually encoded. If you’re going by bike, you’ll probably also check the contour lines to choose a not too steep road.
Pathline also hides structures that may be very useful to biology students, but are well known by the end of their studies. At this point, you know how a certain pathway works and all you need is hints that help you see, which part of a pathway you’re looking at.
Two main advantages of the new tool were found. First, there was a massive gain in efficiency. The study of a heatmap took up to a half-hour before but can be done at a glance at the curvemap now. More importantly, though, the scientists made new discoveries of gene properties they didn’t know about before. What was hidden in the data before, is now very clear, even to an untrained eye.
What we see is that genes g5 and g6 are very similarly expressed across all species – except for species s7. Thanks to this tool, the scientists were able to see this anomaly and to conduct more research into why this is so.
We think Pathline is an excellent example of researchers pushing science forward through visualization. It’s striking how much better the visualizations are now compared to what they were before. From our experience, a collaboration between scientists and designers is very difficult, as design schools are usually not in the same places as more science oriented universities, so we need to find ways to collaborate more. But we hope that by covering a topic like this, readers are encouraged to go out and look for projects in an interdisciplinary scientific context.
Pathline is available as an open source Processing application from the Pathline website – hopefully soon on GitHub as well, so we can contribute back more easily if we see opportunities for improvement.