Using hypeR in Shiny Applications
In hypeR: An R Package For Geneset Enrichment Workflows

knitr::opts_chunk$set(comment="", fig.align="center", fig.width=10, fig.height=8, cache=FALSE)
devtools::load_all(".")

Shiny Modules

All hypeR methods output objects that are compatible with R Shiny and can be incorporated into a variety of applications (e.g. hyp_dots() returns a ggplot object). The most difficult challenge in incorporating geneset enrichment tasks in Shiny applications is the geneselect selection / fetching itself. To help with this, hypeR includes module functions that abstract away the geneset selection code.

What Are Shiny Modules?

Shiny modules are functions that generato Shiny UI / Server code. They are composable elements that can be used within and interface with existing applications. For more information, check out the latest documentation.

Example

For those familiar with Shiny, you will have accompanying modules for the ui and server sections of your application. These modules are hypeR::genesets_UI() and hypeR::genesets_Server() respectively.

UI

Here we create a simple interface where we use the genesets selection module. Through that module, users can select genesets available through hypeR and the application will load those genesets into a reactive variable (more on this later). We add additional code to take a user-defined gene signature and produce an enrichment plot which represents components specific to your Shiny application. This can be anything!

ui <- fluidPage(
        sidebarLayout(
            sidebarPanel(
                # Put your geneset selector module anywhere
                hypeR::genesets_UI("genesets"),

                # Add components specific to your application
                textAreaInput("signature", 
                              label="Signature", 
                              rows=5,
                              placeholder="GENE1,GENE2,GENE3", 
                              resize="vertical"),

                actionButton("enrichment", "Enrichment")
            ),
            mainPanel(
                # Enrichment plot
                plotOutput("plot")
            )
        )
    )

Server

Now we need to create the server code that is responsible for all the backend work. Again, we provide an associated module function for the server code. The serve code returns a reactive variable containing the fetched genesets that were selected. Selection changes will update this variable and propogate to downstream functions. Importantly, this variable holding the genesets can be accessed by any part of your application now, enabling you to make applications completely customizable while still utilizing this feature. As an example, our custom function is producing a simple enrichment plot.

server <- function(input, output, session) {
    # Retrieve selected genesets as a reactive variable
    # Selection changes will update this variable and propogate to downstream functions
    genesets <- hypeR::genesets_Server("genesets")

    # Your custom downstream functions
    reactive_plot <- eventReactive(input$enrichment, {

        # Here are the fetched genesets
        gsets <- genesets()

        # Process the signature into a character vector
        signature <- input$signature %>%
                     stringr::str_split(pattern=",", simplify=TRUE) %>%
                     as.vector()

        # Run hypeR
        hyp <- hypeR::hypeR(signature, gsets, test="hypergeometric")
        p <- hypeR::hyp_dots(hyp, top=10, fdr=0.25)

        # These are just ggplot objects you could customize
        p + theme(axis.text=element_text(size=12, face="bold"))
    })

    output$plot <- renderPlot({
        reactive_plot()
    })
}