server-boxplot.R

observe({
  updateSelectizeInput(session,'sel_gene',
                       choices= rownames(myValues$dataCounts),
                       server=TRUE)
  updateSelectizeInput(session,'sel_groups',
                       choices= colnames(myValues$DF),
                       server=TRUE)
})
observe({

  
  updateSelectInput(session,'boxplotX',
                       choices= colnames(myValues$DF),
                    selected = colnames(myValues$DF)[1])
  
  updateSelectInput(session,'boxplotFill',
                       choices= colnames(myValues$DF),
                    selected = colnames(myValues$DF)[1])
  #tmpgroups = unique(myValues$DF$Conditions)

  tmpgroups = input$sel_groups
  tmpgroups = unlist(lapply(tmpgroups, function(x) {levels(myValues$DF[,x])}))
  
  updateSelectizeInput(session,'sel_factors',
                           choices=tmpgroups, selected=tmpgroups, server = T)
  
  
})

observe({
  geneExrReactive()
})

geneExrReactive <- reactive({


      validate(need(length(input$sel_gene)>0,"Please select a gene."))
      validate(need(length(input$sel_groups)>0,"Please select a group(s)."))
      validate(need(length(input$sel_factors)>0,"Please select factors."))
      
      
      filtered <- t(log2((counts(myValues$dds[input$sel_gene, ], normalized=TRUE, replaced=FALSE)+.5))) %>%
           merge(colData(myValues$dds), ., by="row.names") %>%
           gather(gene, expression, (ncol(.)-length(input$sel_gene)+1):ncol(.))
      
      
      factors = input$sel_groups
      
      
      filtered_new = filtered
      for(i in 1:length(factors))
      {
        f = factors[i]
        filtered_new = inner_join(filtered_new, filtered[filtered[,f] %in% input$sel_factors,])
      }
      
      return(filtered_new)
 
    
})

output$boxPlot <- renderPlotly({
  if(!is.null(geneExrReactive()))
  {
    filtered = geneExrReactive()
    
    validate(need(length(input$boxplotX)>0,"Please select a group."))
    validate(need(length(input$boxplotFill)>0,"Please select a fill by group."))
    
    ## Adapted from STARTapp dotplot
    
    p <- ggplot(filtered, aes_string(input$boxplotX, "expression", fill=input$boxplotFill)) + 
      geom_boxplot() + facet_wrap(~gene, scales="free_y")
    
    
    p <- p + xlab(" ") + theme(
      plot.margin = unit(c(1,1,1,1), "cm"),
      axis.text.x = element_text(angle = 45),
      legend.position="bottom")
    
    p
    
  }
  
  
})

output$boxplotData <- renderDataTable({
  if(!is.null(geneExrReactive()))
  {
    geneExrReactive()
  }
}, 
options = list(scrollX = TRUE, pageLength = 5))


output$boxplotComputed <- reactive({
  return(!is.null(geneExrReactive()))
})
outputOptions(output, 'boxplotComputed', suspendWhenHidden=FALSE)


output$downloadBoxCsv <- downloadHandler(

  filename = function() {paste0(input$boxplotX,"_",input$boxplotFill,"_boxplotdata.csv")},
  content = function(file) {
    csv = geneExrReactive()

    write.csv(csv, file, row.names=F)
  }

)