Fig. 2: Workflow of the PECAn R Shiny statistical analysis application.

a Users upload the datasets generated by the FIJI/ImageJ plugin (1) and group images by experiment and treatment/genotype/condition using the graphical user interface (2). The user can then assess their samples either via classical uni- and bi-variate tests (a/blue) or via multiple regression analysis (b/yellow). To perform uni- and bi-variate tests, the user specifies an analysis to perform (3a) and then selects an appropriate plot to generate, which tests to run, which reference group to use, and which (if any) p-correction for multiple comparisons to perform (4a). Upon running the analysis (5a), output graphs and statistical tests are performed, along with tests for parametric assumptions. To perform multiple regression analysis, the user specifies which form of regression to run and which parameter should act as the dependent variable (3b). The user then selects which parameters to include as predictor variables (4b), which test to use for generating a correlation matrix, and which (if any) p-correction for multiple comparisons to perform (5b). Upon running the analysis (6b), the regression is run along with test for assumptions and both output and diagnostic plots are generated. b A summary of the kinds of tests supported by the app. c Examples of plots generated by the app, using the ggplot2 package in R. Different plots are generated for uni- and bi-variate tests (left/blue) and for multiple regression analysis (right/yellow).