user2016

We all know the adage “A picture is worth a thousand words,” but when it comes to multi-dimensional data, there are challenges to maintaining this visual brevity; this case study provides an example. A dose combination experiment was conducted wherein cytokine expression in T cells was measured in response to two compounds administered in combination across a set range of doses. Response surface regression was used to find the dose combination with the highest level of response. Biological response to this ‘best’ dose combination was then compared against response to single-dose comparator compound. Tables of summary statistics, conditional boxplots, snapshots of 3D scatterplots, and paragraphs describing the first derivative are all perfectly adequate to present results from this experiment in a manuscript. However for fast communication of complicated results, a Shiny app succeeds admirably. This particular Shiny app, which utilizes the shiny, rgl, and shinyRGL packages, provides a dynamic and flexible solution to the difficulty of displaying literally all sides of this multi-dimensional analysis to live audiences. Rather than having to sit through an explanation of local extrema versus saddle points, scientist can quickly generate the shape of biological response over dose combinations and examine how the best dose combination stacks up against the comparator. Moreover, users can also take advantage of the portability that deployment on a Shiny server or Shinyapps.io provides to those without R access, allowing them to present results without requiring additional software or a spare statistician.