inst/shiny_apps/LFQ/app.R

In DEP: Differential Enrichment analysis of Proteomics data

library(dplyr)
library(tibble)
library(SummarizedExperiment)
library(DEP)
library(shiny)
library(shinydashboard)

ui <- shinyUI(
  dashboardPage(
    dashboardHeader(title = "DEP - LFQ"),
    dashboardSidebar(
      sidebarMenu(
        menuItem("Files", selected = TRUE,
                 fileInput('file1',
                           'ProteinGroups.txt',
                           accept=c('text/csv',
                                    'text/comma-separated-values,text/plain',
                                    '.csv')),
                 fileInput('file2',
                           'ExperimentalDesign.txt',
                           accept=c('text/csv',
                                    'text/comma-separated-values,text/plain',
                                    '.csv')),
                 radioButtons("anno",
                              "Sample annotation",
                              choices = list("Parse from columns" = "columns",
                                             "Use Experimental Design" = "expdesign"),
                              selected = "expdesign")
        ),
        menuItemOutput("columns"),
        menuItem("Imputation options",
                 radioButtons("imputation",
                              "Imputation type",
                              choices = c("man", MSnbase::imputeMethods())[1:9],
                              selected = "MinProb"),
                 p(a("Detailed information link",
                     href = "https://www.rdocumentation.org/packages/MSnbase/versions/1.20.7/topics/impute-methods",
                     target="_blank"))
        ),
        actionButton("analyze", "Analyze"),
        tags$hr(),
        uiOutput("downloadTable"),
        uiOutput("downloadButton")
      )
    ),
    dashboardBody(
      helpText("Please cite: "),
      fluidRow(
        box(numericInput("p",
                         "adj. P value",
                         min = 0.0001, max = 0.1, value = 0.05),
            width = 2),
        box(numericInput("lfc",
                         "Log2 fold change",
                         min = 0, max = 10, value = 1),
            width = 2),
        infoBoxOutput("significantBox"),
        box(radioButtons("pres",
                         "Data presentation",
                         c("contrast", "centered"),
                         selected = "contrast"),
            width = 2),
        box(radioButtons("contrasts",
                         "Contrasts",
                         c("control", "all"),
                         selected = "control"),
            width = 2)
      ),
      fluidRow(
        column(width = 7,
               box(title = "Top Table",
                   box(uiOutput("select"), width = 6),
                   box(uiOutput("exclude"), width = 6),
                   DT::dataTableOutput("table"), width = 12)
        ),
        column(width = 5,
               tabBox(title = "Result Plots", width = 12,
                      tabPanel(title = "Selected Protein",
                               plotOutput("selected_plot"),
                               downloadButton('downloadPlot', 'Save plot')),
                      tabPanel(title = "Heatmap",
                               fluidRow(
                                 box(numericInput("k",
                                                  "Kmeans clusters",
                                                  min = 0, max = 15, value = 7),
                                     width = 4),
                                 box(numericInput("limit",
                                                  "Color limit (log2)",
                                                  min = 0, max = 16, value = 6),
                                     width = 4),
                                 box(numericInput("size",
                                                  "Heatmap size (4-30)",
                                                  min = 4, max = 30, value = 10),
                                     width = 4)
                               ),
                               fluidRow(
                                 uiOutput("plot"),
                                 downloadButton('downloadHeatmap', 'Save heatmap'))
                      ),
                      tabPanel(title = "Volcano plot",
                               fluidRow(
                                 box(uiOutput("volcano_cntrst"), width = 6),
                                 box(numericInput("fontsize",
                                                  "Font size",
                                                  min = 0, max = 8, value = 4),
                                     width = 3),
                                 box(checkboxInput("check_names",
                                                   "Display names",
                                                   value = TRUE),
                                     checkboxInput("p_adj",
                                                   "Adjusted p values",
                                                   value = FALSE),
                                     width = 3)
                               ),
                               fluidRow(
                                 plotOutput("volcano", height = 600),
                                 downloadButton('downloadVolcano', 'Save volcano')
                               )
                      )
               ),
               tabBox(title = "QC Plots", width = 12,
                      tabPanel(title = "Protein Numbers",
                               plotOutput("numbers", height = 600),
                               downloadButton('downloadNumbers', 'Save')
                      ),
                      tabPanel(title = "Sample coverage",
                               plotOutput("coverage", height = 600),
                               downloadButton('downloadCoverage', 'Save')
                      ),
                      tabPanel(title = "Normalization",
                               plotOutput("norm", height = 600),
                               downloadButton('downloadNorm', 'Save')
                      ),
                      tabPanel(title = "Missing values - Quant",
                               plotOutput("detect", height = 600),
                               downloadButton('downloadDetect', 'Save')
                      ),
                      tabPanel(title = "Missing values - Heatmap",
                               plotOutput("missval", height = 600),
                               downloadButton('downloadMissval', 'Save')
                      ),
                      tabPanel(title = "Imputation",
                               plotOutput("imputation", height = 600),
                               downloadButton('downloadImputation', 'Save')
                      )
               )
        )
      )
    )
  )
)

server <- shinyServer(function(input, output) {
  options(shiny.maxRequestSize=200*1024^2)

  ### UI functions ### --------------------------------------------------------
  output$columns <- renderMenu({
    menuItem("Columns",
             selectizeInput("name",
                            "Name column",
                            choices=colnames(data()),
                            selected = "Gene.names"),
             selectizeInput("id",
                            "ID column",
                            choices=colnames(data()),
                            selected = "Protein.IDs"),
             selectizeInput("filt",
                            "Filter on columns" ,
                            colnames(data()),
                            multiple = TRUE,
                            selected = c("Reverse","Potential.contaminant")),
             if (input$anno == "columns" & !is.null(data())) {
               cols <- grep("^LFQ", colnames(data()))
               prefix <- get_prefix(data()[,cols] %>% colnames())
               selectizeInput("control", "Control",
                              choices=make.names(colnames(data())[cols] %>%
                                                   gsub(prefix,"",.) %>%
                                                   substr(., 1, nchar(.)-1)))
             },
             if (input$anno == "expdesign" & !is.null(expdesign())) {
               selectizeInput("control",
                              "Control",
                              choices = make.names(expdesign()$condition))
             }
    )
  })

  ### Reactive functions ### --------------------------------------------------
  data <- reactive({
    inFile <- input$file1
    if (is.null(inFile))
      return(NULL)
    read.csv(inFile$datapath, header = TRUE,
             sep = "\t", stringsAsFactors = FALSE) %>%
      mutate(id = row_number())
  })
  expdesign <- reactive({
    inFile <- input$file2
    if (is.null(inFile))
      return(NULL)
    read.csv(inFile$datapath, header = TRUE,
             sep = "\t", stringsAsFactors = FALSE) %>%
      mutate(id = row_number())
  })

  filt <- reactive({
    data <- data()
    cols <- grep("^LFQ", colnames(data))

    filtered <- DEP:::filter_MaxQuant(data, input$filt)
    unique_names <- make_unique(filtered, input$name, input$id)

    if (input$anno == "columns") {
      se <- make_se_parse(unique_names, cols)
    }
    if (input$anno == "expdesign") {
      se <- make_se(unique_names, cols, expdesign())
    }
    filter_missval(se, 0)
  })

  norm <- reactive({
    normalize_vsn(filt())
  })

  imp <- reactive({
    DEP::impute(norm(), input$imputation)
  })

  df <- reactive({
    validate(
      need(input$control != "", "Please select a control condition")
    )
    test_diff(imp(), input$contrasts, input$control)
  })

  dep <- reactive({
    add_rejections(df(), input$p, input$lfc)
  })

  ### All object and functions upon 'Analyze' input  ### ----------------------

  observeEvent(input$analyze, {

    ### Interactive UI functions ### ------------------------------------------
    output$downloadTable <- renderUI({
      selectizeInput("dataset",
                     "Choose a dataset to save" ,
                     c("results","significant_proteins",
                       "displayed_subset","full_dataset"))
    })

    output$downloadButton <- renderUI({
      downloadButton('downloadData', 'Save')
    })

    output$significantBox <- renderInfoBox({
      num_total <- dep() %>%
        nrow()
      num_signif <- dep() %>%
        .[rowData(.)$significant, ] %>%
        nrow()
      frac <- num_signif / num_total

      if(frac > 0.2) {
        info_box <- infoBox("Significant proteins",
                            paste0(num_signif,
                                   " out of ",
                                   num_total),
                            paste0("Too large fraction (",
                                   signif(frac * 100, digits = 3),
                                   "%) of proteins differentially expressed"),
                            icon = icon("minus", lib = "glyphicon"),
                            color = "orange",
                            width = 4)
      }
      if(frac == 0) {
        info_box <- infoBox("Significant proteins",
                            paste0(num_signif,
                                   " out of ",
                                   num_total),
                            "No proteins differentially expressed",
                            icon = icon("thumbs-down", lib = "glyphicon"),
                            color = "red",
                            width = 4)
      }
      if(frac > 0 & frac <= 0.2) {
        info_box <- 		infoBox("Significant proteins",
                              paste0(num_signif,
                                     " out of ",
                                     num_total),
                              paste0(signif(frac * 100, digits = 3),
                                     "% of proteins differentially expressed"),
                              icon = icon("thumbs-up", lib = "glyphicon"),
                              color = "green",
                              width = 4)
      }
      info_box
    })

    output$select <- renderUI({
      row_data <- rowData(dep())
      cols <- grep("_significant", colnames(row_data))
      names <- colnames(row_data)[cols]
      names <- gsub("_significant", "", names)
      selectizeInput("select",
                     "Select direct comparisons",
                     choices=names,
                     multiple = TRUE)
    })

    output$exclude <- renderUI({
      row_data <- rowData(dep())
      cols <- grep("_significant", colnames(row_data))
      names <- colnames(row_data)[cols]
      names <- gsub("_significant","",names)
      selectizeInput("exclude",
                     "Exclude direct comparisons",
                     choices = names,
                     multiple = TRUE)
    })

    output$volcano_cntrst <- renderUI({
      if (!is.null(selected())) {
        df <- rowData(selected())
        cols <- grep("_significant$",colnames(df))
        selectizeInput("volcano_cntrst",
                       "Contrast",
                       choices = gsub("_significant", "", colnames(df)[cols]))
      }
    })

    ### Reactive functions ### ------------------------------------------------
    excluded <- reactive({
      DEP:::exclude_deps(dep(), input$exclude)
    })

    selected <- reactive({
      DEP:::select_deps(excluded(), input$select)
    })

    res <- reactive({
      get_results(selected())
    })

    table <- reactive({
      DEP:::get_table(res(), input$pres)
    })

    selected_plot_input <- reactive ({
      if(!is.null(input$table_rows_selected)) {
        selected_id <- table()[input$table_rows_selected,1]
        plot_single(selected(), selected_id, input$pres)
      }
    })

    heatmap_input <- reactive({
      withProgress(message = 'Plotting', value = 0.66, {
        plot_heatmap(selected(),
                     input$pres,
                     kmeans = TRUE,
                     input$k,
                     input$limit)
      })
    })

    volcano_input <- reactive({
      if(!is.null(input$volcano_cntrst)) {
        plot_volcano(selected(),
                     input$volcano_cntrst,
                     input$fontsize,
                     input$check_names,
                     input$p_adj)
      }
    })

    norm_input <- reactive({
      plot_normalization(filt(),
                         norm())
    })

    missval_input <- reactive({
      plot_missval(norm())
    })

    detect_input <- reactive({
      plot_detect(norm())
    })

    imputation_input <- reactive({
      plot_imputation(norm(),
                      df())
    })

    numbers_input <- reactive({
      plot_numbers(norm())
    })

    coverage_input <- reactive({
      plot_coverage(norm())
    })

    ### Output functions ### --------------------------------------------------
    output$table <- DT::renderDataTable({
      table()
    }, options = list(pageLength = 25, scrollX = T),
    selection = list(selected = c(1)))

    output$selected_plot <- renderPlot({
      selected_plot_input()
    })

    output$heatmap <- renderPlot({
      heatmap_input()
    })

    output$volcano <- renderPlot({
      volcano_input()
    })

    output$norm <- renderPlot({
      norm_input()
    })

    output$missval <- renderPlot({
      missval_input()
    })

    output$detect <- renderPlot({
      detect_input()
    })

    output$imputation <- renderPlot({
      imputation_input()
    })

    output$numbers <- renderPlot({
      numbers_input()
    })

    output$coverage <- renderPlot({
      coverage_input()
    })

    observe({
      output$plot <- renderUI({
        plotOutput("heatmap", height = (100 * as.numeric(input$size)))
      })
    })

    ### Download objects and functions ### ------------------------------------
    datasetInput <- reactive({
      switch(input$dataset,
             "results" = get_results(dep()),
             "significant_proteins" = get_results(dep()) %>%
               filter(significant) %>%
               select(-significant),
             "displayed_subset" = res() %>%
               filter(significant) %>%
               select(-significant),
             "full_dataset" = get_df_wide(dep()))
    })

    output$downloadData <- downloadHandler(
      filename = function() { paste(input$dataset, ".txt", sep = "") },
      content = function(file) {
        write.table(datasetInput(),
        file,
        col.names = TRUE,
        row.names = FALSE,
        sep ="\t") }
    )

    output$downloadPlot <- downloadHandler(
      filename = function() {
        paste0("Barplot_", table()[input$table_rows_selected,1], ".pdf")
      },
      content = function(file) {
        pdf(file)
        print(selected_plot_input())
        dev.off()
      }
    )

    output$downloadHeatmap <- downloadHandler(
      filename = 'Heatmap.pdf',
      content = function(file) {
        pdf(file, height = 10, paper = "a4")
        print(heatmap_input())
        dev.off()
      }
    )

    output$downloadVolcano <- downloadHandler(
      filename = function() {
        paste0("Volcano_", input$volcano_cntrst, ".pdf")
      },
      content = function(file) {
        pdf(file)
        print(volcano_input())
        dev.off()
      }
    )

    output$downloadNorm <- downloadHandler(
      filename = "normalization.pdf",
      content = function(file) {
        pdf(file)
        print(norm_input())
        dev.off()
      }
    )

    output$downloadMissval <- downloadHandler(
      filename = "missing_values_heatmap.pdf",
      content = function(file) {
        pdf(file)
        print(missval_input())
        dev.off()
      }
    )

    output$downloadDetect <- downloadHandler(
      filename = "missing_values_quant.pdf",
      content = function(file) {
        pdf(file)
        gridExtra::grid.arrange(detect_input())
        dev.off()
      }
    )

    output$downloadImputation <- downloadHandler(
      filename = "imputation.pdf",
      content = function(file) {
        pdf(file)
        print(imputation_input())
        dev.off()
      }
    )

    output$downloadNumbers <- downloadHandler(
      filename = "numbers.pdf",
      content = function(file) {
        pdf(file)
        print(numbers_input())
        dev.off()
      }
    )

    output$downloadCoverage <- downloadHandler(
      filename = "coverage.pdf",
      content = function(file) {
        pdf(file)
        print(coverage_input())
        dev.off()
      }
    )
  })
})


shinyApp(ui = ui, server = server)