inst/HaDeX/server.R

In HaDeX: Analysis and Visualisation of Hydrogen/Deuterium Exchange Mass Spectrometry Data

source("ui.R")



#########################################

server <- function(input, output, session) {
  
  ##
  
  observe_helpers(help_dir = "docs", withMathJax = TRUE)
  
  ### TAB: START ###
  
  ##
  
  output[["file_req"]] <- renderTable({
    
    file_req
    
  })
  
  dat_in <- reactive({
    
    inFile <- input[["data_file"]]
    
    if (is.null(inFile)){
      read_hdx('./data/KD_180110_CD160_HVEM.csv')
    } else {
      validate(need(try(read_hdx(inFile[["datapath"]])), "Check file requirements!"))
      read_hdx(inFile[["datapath"]])
    }
    
  })
  
  output[["data_file_info"]] <- renderText({
    
    if (is.null(input[["data_file"]])){
      "Example file: KD_180110_CD160_HVEM.csv"
    } else {
      length(dat_in()[[1]])
      "Supplied file is valid."
    }
    
  })
  
  ##
  
  ### TAB: INPUT DATA
  
  ##
  
  dat_tmp <- reactive({
    
    dat_in() %>%
      mutate(Start = Start + input[["sequence_start_shift"]] -1,
             End = End + input[["sequence_start_shift"]] -1)
    
  })
  
  ## mark for modifications
  
  has_modifications <- reactive({
    
    any(!is.na(dat_tmp()[["Modification"]]))
    
  })
  
  ##
  
  observe({
    
    if(has_modifications()){
      hide("chosen_control")
    }
    
  })
  
  ##
  
  observe({
    
    if(!has_modifications()){
      show("chosen_control")
    }
    
  })
  
  ##
  
  proteins_from_file <- reactive({
    
    unique(dat_in()[["Protein"]])
    
  })
  
  ##
  
  max_range <- reactive({
    
    max(filter(dat_in(), Protein == input[["chosen_protein"]])[['End']])
    
  })
  
  ##
  
  observe({
    
    updateSelectInput(session,
                      inputId = "chosen_protein",
                      choices = proteins_from_file(),
                      selected = proteins_from_file()[1])
    
  })
  
  ##
  
  options_for_control <- reactive({
    
    dat_in() %>%
      filter(Protein == input[["chosen_protein"]]) %>%
      mutate(Exposure = round(Exposure, 3)) %>%
      select(Protein, State, Exposure) %>%
      arrange(State, desc(Exposure)) %>%
      unique(.) %>%
      mutate(control = paste0(Protein, " | ", State, " | ", Exposure)) %>%
      select(control) 
    
  })
  
  ##
  
  observe({
    
    tryCatch({
      if(input[["deut_concentration"]] < 0)
        updateNumericInput(session,
                           inputId = "deut_concentration",
                           value = 0)
    }, error = function(e){
        updateNumericInput(session,
                           inputId = "deut_concentration",
                           value = 0)
    })
  })
  
  observe({
    
    tryCatch({
      if(input[["deut_concentration"]] > 100)
        updateNumericInput(session,
                           inputId = "deut_concentration",
                           value = 100)
    },
    error = function(e){
        updateNumericInput(session,
                           inputId = "deut_concentration",
                           value = 100)
    })
    
  })
  
  observe({
    
    tryCatch({
      if(input[["sequence_length"]] < max_range())
        updateNumericInput(session,
                           inputId = "sequence_length",
                           value = max_range())
    },
    error = function(e){
        updateNumericInput(session,
                           inputId = "sequence_length",
                           value = max_range())
    })
    
  })
  
  observe({

    tryCatch(
    { if(input[["sequence_start_shift"]] < 0)
        updateNumericInput(session,
                           inputId = "sequence_start_shift",
                           value = 1) },
      error = function(e) {
        message("You cannot shift it to minus values!")
        updateNumericInput(session,
                           inputId = "sequence_start_shift",
                           value = 1) })
  })
  
  ##
  
  observe({
    
    updateSelectInput(session, 
                      inputId = "chosen_control", 
                      choices = options_for_control())
    
    updateNumericInput(session, 
                       inputId = "sequence_length",
                       value = max_range())
    
  })
  
  ##
  
  ## create dat based on control
  
  dat <- reactive({
    
    tmp <- dat_tmp() %>%
      filter(Protein == input[["chosen_protein"]], 
             State == strsplit(input[["chosen_control"]], " \\| ")[[1]][2], 
             Exposure == strsplit(input[["chosen_control"]], " \\| ")[[1]][3]) %>%
      mutate(Exposure = 99999)
    
    states_to_prepare <- unique(filter(dat_tmp(), Protein == input[["chosen_protein"]])[["State"]])
    
    bind_rows(dat_tmp(), 
              lapply(states_to_prepare, function(state){
                peps <- dat_tmp() %>%
                  filter(State == state) %>%
                  select(Sequence) %>%
                  unique(.) %>%
                  unlist(.)
                tmp %>%
                  filter(Sequence %in% peps) %>%
                  mutate(State = state) 
              }))
  })
  
  ##
  
  ### TAB: SEQUENCE DATA ###
  
  ##
  
  observeEvent(input[["data_file"]], {
    
    possible_states <- unique(dat()[["State"]])
    
    updateRadioButtons(session,
                       inputId = "chosen_state",
                       choices = possible_states)
    
  })
  
  ##
  
  protein_name <- reactive({
    
    as.character(unique(dat()[["Protein"]]))
    
  })
  
  ##
  
  output[["protein_name"]] <- renderText({
    
    input[["chosen_protein"]]
    
  })
  
  ##
  
  position_in_sequence_tmp <- reactive({
    
    dat() %>%
      filter(Protein == input[["chosen_protein"]]) %>%
      select(Start, End, Sequence) %>%
      unique(.) %>%
      apply(1, function(x) data.frame(position = x[1]:x[2], amino = strsplit(x[3], '')[[1]], stringsAsFactors = FALSE)) %>%
      bind_rows() %>%
      unique(.) 
    
  })
  
  ##
  
  protein_sequence <- reactive({
    
    reconstruct_sequence(filter(dat(), Protein == input[["chosen_protein"]]))
    
  })
  
  ##
  
  position_in_sequence <- reactive({
    
    position_in_sequence_tmp() %>%
      left_join(amino_prop)
    
  })
  
  ##
  
  output[["protein_stats"]] <- renderTable({
    
    data.frame(
      Name = c("Length", "Coverage", "Cys"),
      Value = as.character(c(input[["sequence_length"]], 
                             paste0(100*round((max_range()-str_count(protein_sequence(), 'x'))/input[["sequence_length"]], 4), '%'),
                             str_count(protein_sequence(), 'C'))),
      stringsAsFactors = FALSE
    )
    
  })
  
  ##
  
  output[["sequence_length_exp_info"]] <- renderText({
    
    paste("Sequence length from the file is ", max_range(), ".")
    
  })
  
  ##
  
  protein_sequence_colored <- reactive({
    
    paste0("<span>", 
           gsubfn(pattern = 'C', replacement = function(x) paste0('<font color = "red">', x, "</font>"), x = protein_sequence()),
           "</span>")
    
  })
  
  ##
  
  output[["sequenceName"]] <- renderText({
    
    protein_sequence_colored()
    
  })
  
  ##
  
  aminoDist_out <- reactive({
    
    charge_colors <- c("-1" = "#E41A1C", "0" = "#377EB8", "1" = "#4DAF4A")
    
    position_in_sequence() %>%
      mutate(affinity = ifelse(is_hydrophobic, "phobic", "philic")) %>% 
      filter(affinity %in% input[["hydro_prop"]]) %>%
      mutate(amino = factor(amino, levels = amino_groups)) %>%
      group_by(amino, charge, is_hydrophobic) %>%
      summarise(cnt = n()) %>%
      ggplot(aes(x = amino, y = cnt, fill = charge)) + 
      geom_col() +
      scale_fill_manual("Charge", values = charge_colors) + 
      labs(title = paste0('Amino acid composition for ', input[["chosen_protein"]]),
           x = 'Amino acid',
           y = 'Count')
    
  })
  
  ##
  
  output[["aminoDist"]] <- renderPlot({
    
    aminoDist_out()
    
  })
  
  ##
  
  output[["aminoDist_debug"]] <- renderUI({
    
    if(!is.null(input[["aminoDist_hover"]])) {
      
      plot_data <- aminoDist_out()[["data"]] %>%
        ungroup()
      
      hv <- input[["aminoDist_hover"]]
      
      hv_dat <- data.frame(x = hv[["x"]],
                           y = hv[["y"]],
                           x_plot = plot_data[[".group"]],
                           y_plot = plot_data[[hv[["mapping"]][["y"]]]],
                           amino = plot_data[["amino"]],
                           charge = plot_data[["charge"]],
                           is_hydrophobic = plot_data[["is_hydrophobic"]],
                           count = plot_data[["cnt"]])
      
      tt_df <- filter(hv_dat, abs(x_plot - x) < 0.5) %>%
        filter(abs(x_plot -x ) == min(abs(x_plot - x)))
      
      if(nrow(tt_df) != 0) { 
        
        tt_pos_adj <- ifelse(hv[["coords_img"]][["x"]]/hv[["range"]][["right"]] < 0.5,
                             "left", "right")
        
        tt_pos <- ifelse(hv[["coords_img"]][["x"]]/hv[["range"]][["right"]] < 0.5,
                         hv[["coords_css"]][["x"]], 
                         hv[["range"]][["right"]]/hv[["img_css_ratio"]][["x"]] - hv[["coords_css"]][["x"]])
        
        
        style <- paste0("position:absolute; z-index:1000; background-color: rgba(245, 245, 245, 1); pointer-events: none;",
                        tt_pos_adj, ":", tt_pos, 
                        "px; top:", hv[["coords_css"]][["y"]], "px; padding: 0px;")
        
        div(
          style = style,
          p(HTML(paste0("<br/> Amino acid: ", tt_df[["amino"]],
                        "<br/> Charge: ", tt_df[["charge"]],
                        "<br/> Is hydrophobic? ", tt_df[["is_hydrophobic"]],
                        "<br/> Count: ", tt_df[["count"]])))
        )
      }
    }
  })
  
  ##
  
  output[["aminoDist_download_button"]] <- downloadHandler("aminoDist.svg",
                                                           content = function(file){
                                                             ggsave(file, aminoDist_out(), device = svg, 
                                                                    height = 300, width = 400, units = "mm")
                                                           })
  ##
  
  ### TAB: COVERAGE ###
  
  ##
  
  stateOverlap_data <- reactive({
    
    dat() %>%
      select(Protein, Sequence, Start, End, State) %>% 
      filter(Protein == input[["chosen_protein"]]) %>%
      filter(State == input[["chosen_state"]]) %>%
      filter(Start >= input[["plot_range"]][[1]], End <= input[["plot_range"]][[2]]) %>%
      filter(!duplicated(.)) %>%
      select(-State)
      
    
  })
  
  output[["stateOverlap_data"]] <- DT::renderDataTable(server = FALSE, {
    
    stateOverlap_data() %>%
      dt_format(cols = c("Protein", "Sequence", "Start", "End"))
    
  })
  
  ##
  
  stateOverlap_out <- reactive({
    
    stateOverlap_data() %>%
      select(Sequence, Start, End) %>%
      filter(!duplicated(.)) %>%
      arrange(Start, End) %>%
      mutate(ID = row_number()) %>%
      ggplot() +
      geom_segment(aes(x = Start, y = ID, xend = End, yend = ID)) +
      labs(title = "Peptide coverage",
           x = "Position",
           y = "") +
      theme(axis.ticks.y = element_blank(),
            axis.text.y = element_blank()) +
      coord_cartesian(xlim = c(input[["plot_range"]][[1]], input[["plot_range"]][[2]]))
    
  })
  
  ##
  
  output[["stateOverlap"]] <- renderPlot({
    
    stateOverlap_out() +
      labs(title = paste0("Peptide coverage for ", input[["chosen_protein"]]))
    
  })
  
  output[["stateOverlap_debug"]] <- renderUI({
    
    if(!is.null(input[["stateOverlap_hover"]])) {
      
      plot_data <- stateOverlap_out()[["data"]]
      hv <- input[["stateOverlap_hover"]]
      
      hv_dat <- data.frame(x = hv[["x"]],
                           y = hv[["y"]],
                           Start = plot_data[["Start"]],
                           End = plot_data[["End"]],
                           y_plot = plot_data[[hv[["mapping"]][["y"]]]],
                           Sequence = plot_data[["Sequence"]])
      
      tt_df <- filter(hv_dat, Start < x, End > x) %>% 
        filter(abs(y_plot - y) < 5) %>%
        filter(abs(y_plot - y) == min(abs(y_plot - y))) 
      
      if(nrow(tt_df) != 0) { 
        
        tt_pos_adj <- ifelse(hv[["coords_img"]][["x"]]/hv[["range"]][["right"]] < 0.5,
                             "left", "right")
        
        tt_pos <- ifelse(hv[["coords_img"]][["x"]]/hv[["range"]][["right"]] < 0.5,
                         hv[["coords_css"]][["x"]], 
                         hv[["range"]][["right"]]/hv[["img_css_ratio"]][["x"]] - hv[["coords_css"]][["x"]])
        
        
        style <- paste0("position:absolute; z-index:1000; background-color: rgba(245, 245, 245, 1); pointer-events: none;",
                        tt_pos_adj, ":", tt_pos, 
                        "px; top:", hv[["coords_css"]][["y"]], "px; padding: 0px;")
        
        div(
          style = style,
          p(HTML(paste0(tt_df[["Sequence"]], 
                        "<br/> Position: ", tt_df[["Start"]], "-", tt_df[["End"]])))
        )
      }
    }
  })
  
  ##
  
  ##
  
  output[["stateOverlap_download_button"]] <- downloadHandler("stateOverlap.svg",
                                                              content = function(file){
                                                                ggsave(file, stateOverlap_out(), device = svg, 
                                                                       height = 300, width = 400, units = "mm")
                                                              })
  
  ##
  
  stateOverlapDist_data <- reactive({
    
    dat() %>%
      select(Protein, Start, End, State, Sequence) %>%
      filter(Protein == input[["chosen_protein"]]) %>%
      filter(State == input[["chosen_state"]]) %>%
      filter(Start >= input[["plot_range"]][[1]], End <= input[["plot_range"]][[2]]) %>%
      filter(!duplicated(.)) %>%
      select(-State, -Protein) %>%
      apply(1, function(i) i[1]:i[2]) %>%
      unlist %>%
      data.frame(pos = .) %>%
      group_by(pos) %>%
      summarise(coverage = length(pos)) %>%
      right_join(data.frame(pos = seq(from = input[["plot_range"]][[1]], to = input[["plot_range"]][[2]]))) %>%
      replace_na(list(coverage = 0)) %>%
      right_join(data.frame(amino = unlist(strsplit(protein_sequence(), "")), 
                            pos = 1:str_length(protein_sequence()))) %>%
      select(pos, amino, coverage)
    
  })
  
  ##
  
  output[["stateOverlapDist_data"]] <- DT::renderDataTable(server = FALSE, {
    
    dt_format(stateOverlapDist_data(),
              cols = c("Position", "Amino acid", "Coverage"))
    
  })
  
  ##
  
  stateOverlapDist <- reactive({
    
    mean_coverage <- round(mean(stateOverlapDist_data()[["coverage"]], na.rm = TRUE), 2)
    display_position <- (input[["plot_range"]][[1]] + input[["plot_range"]][[2]])/2
    
    stateOverlapDist_data() %>% 
      ggplot(aes(x = pos, y = coverage)) +
      geom_col(width = 1) +
      labs(x = 'Position', y = 'Position frequency in peptides') +
      theme(legend.position = "none") + 
      coord_cartesian(xlim = c(input[["plot_range"]][[1]], input[["plot_range"]][[2]])) +
      geom_hline(yintercept = mean_coverage, color = 'red') +
      geom_text(aes(x = display_position, y = mean_coverage), label = paste0("Average frequency: ", mean_coverage), color = 'red', vjust = -.5)
    
    
  })
  
  ##
  
  output[["stateOverlapDist"]] <- renderPlot({
    
    stateOverlapDist()
    
  })
  
  ##
  
  output[["stateOverlapDist_debug"]] <- renderUI({
    
    if(!is.null(input[["stateOverlapDist_hover"]])) {
      
      plot_data <- stateOverlapDist()[["data"]]
      hv <- input[["stateOverlapDist_hover"]]
      
      hv_dat <- data.frame(x = hv[["x"]],
                           y = hv[["y"]],
                           x_plot = plot_data[["pos"]],
                           amino = plot_data[["amino"]],
                           coverage = plot_data[["coverage"]])
      
      tt_df <- filter(hv_dat, abs(x - x_plot) < 0.5)  
      
      if(nrow(tt_df) != 0) { 
        
        tt_pos_adj <- ifelse(hv[["coords_img"]][["x"]]/hv[["range"]][["right"]] < 0.5,
                                "left", "right")
        
        tt_pos <- ifelse(hv[["coords_img"]][["x"]]/hv[["range"]][["right"]] < 0.5,
                         hv[["coords_css"]][["x"]], 
                         hv[["range"]][["right"]]/hv[["img_css_ratio"]][["x"]] - hv[["coords_css"]][["x"]])
        
        
        style <- paste0("position:absolute; z-index:1000; background-color: rgba(245, 245, 245, 1); pointer-events: none;",
                        tt_pos_adj, ":", tt_pos, "px; padding: 0px;",
                        "top:", hv[["coords_css"]][["y"]] , "px; ") 
        
        div(
          style = style,
          p(HTML(paste0("<br/> Position: ", tt_df[["x_plot"]],
                        "<br/> Amino acid: ", tt_df[["amino"]], 
                        "<br/> Coverage: ", tt_df[["coverage"]])))
        )
      }
    }
  })
  
  ##
 
  output[["stateOverlapDist_download_button"]] <- downloadHandler("stateOverlapDist.svg",
                                                                  content = function(file){
                                                                    ggsave(file, stateOverlapDist(), device = svg,
                                                                           height = 300, width = 400, units = "mm")
                                                                  })
  ##
  
  ### TAB: WOODS PLOT ###
  
  ##
  
  output[["protein_length"]] <- renderText({
    
    max_range()
    
  })
  
  ##
  
  states_from_file <- reactive({
    
    unique(dat()[["State"]])
    
  })
  
  ## modification actions
  
  observe({
    
    times_from_file <- unique(round(dat()[["Exposure"]], 3))
    
    tmp <- unique(round(dat()[["Exposure"]], 3))
    choose_time_out <- setNames(tmp, c(head(tmp, -1), "chosen control"))
    
    if(has_modifications()){
      
      updateSelectInput(session, 
                        inputId = "out_time",
                        choices = times_from_file[times_from_file < 99999],
                        selected = max(times_from_file[times_from_file < 99999]))
      
    }
    
    if(!has_modifications()){
      
      updateSelectInput(session, 
                        inputId = "out_time",
                        choices = choose_time_out,
                        selected = choose_time_out["chosen control"])
    }
  })
  
  ##
  
  observe({
    
    times_from_file <- unique(round(dat()[["Exposure"]], 3))
    
    tmp <- unique(round(dat()[["Exposure"]], 3))
    choose_time_out <- setNames(tmp, c(head(tmp, -1), "chosen control"))
    
    updateSelectInput(session, 
                      inputId = "chosen_time",
                      choices = times_from_file[times_from_file < 99999],
                      selected = min(times_from_file[times_from_file >= 1]))
    
    updateSelectInput(session, 
                      inputId = "in_time",
                      choices = times_from_file[times_from_file < 99999],
                      selected = min(times_from_file[times_from_file > 0]))
    
    updateSelectInput(session,
                      inputId = "state_first",
                      choices = states_from_file(),
                      selected = states_from_file()[1])
    
    updateSelectInput(session,
                      inputId = "state_second",
                      choices = states_from_file(),
                      selected = states_from_file()[length(states_from_file())])
    
    updateCheckboxGroupInput(session,
                             inputId = "compare_states",
                             choices = states_from_file(),
                             selected = states_from_file())
    
    updateSelectInput(session,
                      inputId = "confidence_limit_2",
                      choices = confidence_limit_choices[confidence_limit_choices >= input[["confidence_limit"]]],
                      selected = confidence_limit_choices[confidence_limit_choices > input[["confidence_limit"]]][1])
    
    updateSliderInput(session, 
                      inputId = "plot_range",
                      max = max_range(),
                      value = c(1, max_range()))
    
    updateSliderInput(session, 
                      inputId = "plot_x_range",
                      max = max_range(),
                      value = c(1, max_range()))
    
  })
  
  ##
  
  observe({
    
    if (input[["calc_type"]] == "absolute") {
      
      min_comparison_abs <- round_any(min(prep_dat()[c("abs_frac_exch_state", "abs_avg_theo_in_time")], na.rm = TRUE), 5, floor)
      max_comparison_abs <- round_any(max(prep_dat()[c("abs_frac_exch_state", "abs_avg_theo_in_time")], na.rm = TRUE), 5, ceiling)
      
      updateSliderInput(session,
                        inputId = "comp_plot_y_range",
                        min = min_comparison_abs - 5,
                        max = max_comparison_abs + 5,
                        value = c(min_comparison_abs, max_comparison_abs),
                        step = 1)
      
      min_woods_abs <- round_any(min(woods_plot_dat()[c("abs_diff_frac_exch", "abs_diff_theo_frac_exch")], na.rm = TRUE), 2, floor)
      max_woods_abs <- round_any(max(woods_plot_dat()[c("abs_diff_frac_exch", "abs_diff_theo_frac_exch")], na.rm = TRUE), 2, ceiling)
      
      updateSliderInput(session,
                        inputId = "woods_plot_y_range",
                        min = min_woods_abs - 2, 
                        max = max_woods_abs + 2, 
                        value = c(min_woods_abs, max_woods_abs),
                        step = 0.5)
      
    } else {
      
      updateSliderInput(session,
                        inputId = "comp_plot_y_range",
                        min = -200,
                        max = 200,
                        value = c(0, 120),
                        step = 10)
      
      updateSliderInput(session,
                        inputId = "woods_plot_y_range",
                        min = -200, 
                        max = 200, 
                        value = c(-50, 50),
                        step = 10)
    }
    
  })
  
  ##
  
  observe({
    
    updateTextInput(session, 
                    inputId = "comparison_plot_title",
                    value = case_when(
                      input[["theory"]] & input[["calc_type"]] == "relative" ~ paste0("Theoretical fraction exchanged in state comparison in ", input[["chosen_time"]], " min for ", input[["chosen_protein"]]),
                      input[["theory"]] & input[["calc_type"]] == "absolute" ~ paste0("Theoretical absolute value exchanged in state comparison in ", input[["chosen_time"]], " min for ", input[["chosen_protein"]]),
                      !input[["theory"]] & input[["calc_type"]] == "relative" ~ paste0("Fraction exchanged in state comparison in ", input[["chosen_time"]], " min for ", input[["chosen_protein"]]),
                      !input[["theory"]] & input[["calc_type"]] == "absolute" ~ paste0("Absolute value exchanged in state comparison in ", input[["chosen_time"]], " min for ", input[["chosen_protein"]])
                    ))
    
    updateTextInput(session, 
                    inputId = "woods_plot_title",
                    value = case_when(
                      input[["theory"]] & input[["calc_type"]] == "relative" ~ paste0("Delta Theoretical fraction exchanged in ", input[["chosen_time"]], " min between ", gsub("_", " ", input[["state_first"]]), " and ", gsub("_", " ", input[["state_second"]]), " for ", input[["chosen_protein"]]),
                      input[["theory"]] & input[["calc_type"]] == "absolute" ~ paste0("Delta Theoretical fraction exchanged in ", input[["chosen_time"]], " min between ", gsub("_", " ", input[["state_first"]]), " and ", gsub("_", " ", input[["state_second"]]), " for ", input[["chosen_protein"]]),
                      !input[["theory"]] & input[["calc_type"]] == "relative" ~ paste0("Delta Fraction exchanged in ", input[["chosen_time"]], " min between ", gsub("_", " ", input[["state_first"]]), " and ", gsub("_", " ", input[["state_second"]]), " for ", input[["chosen_protein"]]),
                      !input[["theory"]] & input[["calc_type"]] == "absolute" ~ paste0("Delta Fraction exchanged in ", input[["chosen_time"]], " min between ", gsub("_", " ", input[["state_first"]]), " and ", gsub("_", " ", input[["state_second"]]), " for ", input[["chosen_protein"]])
                    ))
    
    updateTextInput(session, 
                    inputId = "comparison_plot_y_label", 
                    value = case_when(
                      input[["theory"]] & input[["calc_type"]] == "relative" ~ "Theoretical fraction exchanged [%]",
                      input[["theory"]] & input[["calc_type"]] == "absolute" ~ "Theoretical absolute value exchanged [Da]",
                      !input[["theory"]] & input[["calc_type"]] == "relative" ~ "Fraction exchanged [%]",
                      !input[["theory"]] & input[["calc_type"]] == "absolute" ~ "Absolute value exchanged [Da]"
                    ))
    
    updateTextInput(session, 
                    inputId = "woods_plot_y_label", 
                    value = case_when(
                      input[["theory"]] & input[["calc_type"]] == "relative" ~ "Delta Theoretical fraction exchanged between states [%]",
                      input[["theory"]] & input[["calc_type"]] == "absolute" ~ "Delta Theoretical absolute value exchanged between states [Da]",
                      !input[["theory"]] & input[["calc_type"]] == "relative" ~ "Delta Fraction exchanged between states [%]",
                      !input[["theory"]] & input[["calc_type"]] == "absolute" ~ "Delta Absolute value exchanged between states [Da]"
                    ))
    
  })
  
  ##
  
  observe({
    
    if(input[["theory"]]){
      hide(id = "in_time_part")
      hide(id = "out_time_part")
    }
    
  })
  
  observe({
    
    if(!input[["theory"]]){
      show(id = "in_time_part")
      show(id = "out_time_part")
    }
    
  })
  
  ##
  
  observe({
    
    if(input[["calc_type"]] == "absolute"){
      hide(id = "out_time_part")
    } 
    
  })
    
  ##
  
  observe({
      
    if(input[["calc_type"]] == "relative" & !input[["theory"]]){
      show(id = "out_time_part")
    }
    
  })
  
  ##
  
  comparison_plot_colors <- reactive({
    
    hcl.colors(length(states_from_file()), palette = "Set 2", alpha = NULL, rev = FALSE, fixup = TRUE)
    
  })
  
  ##
  
  output[["states_colors"]] <- renderUI({
    
    # colorInput <- function(inputId, label, value = "", width = NULL, placeholder = NULL) {
    #   '%BAND%' <- function (x, y) {
    #     if (!is.null(x) && !is.na(x)) 
    #       if (!is.null(y) && !is.na(y)) 
    #         return(y)
    #     return(NULL)
    #   }
    #   value <- restoreInput(id = inputId, default = value)
    #   if(!is.null(value))
    #     div(class = "form-group shiny-input-container", 
    #         style = paste(if (value != "") paste0("background-color=: ", validateCssUnit(width), ";"),
    #                       if (!is.null(width)) paste0("width: ", validateCssUnit(width), ";")
    #         ), 
    #         label %BAND%  
    #         tags$label(label, `for` = inputId), tags$input(id = inputId, 
    #                                                        type = "text", class = "form-control", value = value, 
    #                                                        placeholder = placeholder))
    # }
    
    lapply(1:length(states_from_file()), function(i) {
      textInput(inputId = paste0(states_from_file()[i], "_color"),
                label = paste(states_from_file()[i], " color"),
                value = comparison_plot_colors()[i])
    })
  })
  
  ##
  
  comparison_plot_colors_chosen <- reactive({
    
    lapply(paste0(states_from_file(), "_color"), function(i) input[[i]])
    
    tmp <- t(sapply(paste0(input[["compare_states"]],"_color"), function(i) input[[i]][1], simplify = TRUE))
    
    tmp[tmp == "NULL"] <- NA
    
    if (all(is.na(tmp))) {
      comparison_plot_colors()[1:length(states_from_file())]
    } else {
      coalesce(as.vector(tmp), comparison_plot_colors()[1:length(input[["compare_states"]])])
    }
    
  })
  
  ##
  
  ## COMPARISON PLOT + DATA
  
  ##
  
  all_dat <- reactive({
    
    if(!input[["theory"]]){
      validate(need(as.numeric(input[["in_time"]]) < as.numeric(input[["chosen_time"]]), "In time must be smaller than chosen time."))
      validate(need(as.numeric(input[["chosen_time"]]) < as.numeric(input[["out_time"]]), "Out time must be bigger than chosen time."))
    }
    
    
    
    bind_rows(lapply(states_from_file(), function(i) calculate_state_deuteration(dat(), 
                                                                                 protein = input[["chosen_protein"]], 
                                                                                 state = i, 
                                                                                 time_in = input[["in_time"]],
                                                                                 time_chosen = input[["chosen_time"]], 
                                                                                 time_out = input[["out_time"]],
                                                                                 deut_part = 0.01*as.integer(input[["deut_concentration"]]))))
  })
  
  ##
  
  prep_dat <- reactive({
    
    validate(need(input[["compare_states"]], "Please select at least one state."))
    
    filter(all_dat(), State %in% input[["compare_states"]])
    
  })
  
  
  ##
  
  comparison_plot_theo <- reactive({
    
    ggplot(data = prep_dat()) +
      geom_segment(data = prep_dat(), aes(x = Start, y = avg_theo_in_time, xend = End, yend = avg_theo_in_time, color = State)) +
      geom_errorbar(data = prep_dat(), aes(x = Med_Sequence, ymin = avg_theo_in_time - err_avg_theo_in_time, ymax = avg_theo_in_time + err_avg_theo_in_time, color = State)) +
      theme(legend.position = "bottom",
            legend.title = element_blank()) +
      scale_y_continuous(breaks = seq(-200, 200, 10), expand = c(0, 0))
    
  })
  
  ##
  
  comparison_plot_theo_abs <- reactive({
    
    ggplot(data = prep_dat()) +
      geom_segment(data = prep_dat(), aes(x = Start, y = abs_avg_theo_in_time, xend = End, yend = abs_avg_theo_in_time, color = State)) +
      geom_errorbar(data = prep_dat(), aes(x = Med_Sequence, ymin = abs_avg_theo_in_time - err_abs_avg_theo_in_time, ymax = abs_avg_theo_in_time + err_abs_avg_theo_in_time, color = State)) +
      theme(legend.position = "bottom",
            legend.title = element_blank()) +
      scale_y_continuous(expand = c(0, 0))
    
  })
  
  ##
  
  comparison_plot_exp <- reactive({
    
    ggplot(data = prep_dat()) +
      geom_segment(data = prep_dat(), aes(x = Start, y = frac_exch_state, xend = End, yend = frac_exch_state, color = State)) +
      geom_errorbar(data = prep_dat(), aes(x = Med_Sequence, ymin = frac_exch_state - err_frac_exch_state, ymax = frac_exch_state + err_frac_exch_state, color = State)) +
      theme(legend.position = "bottom",
            legend.title = element_blank()) +
      scale_y_continuous(breaks = seq(-200, 200, 10), expand = c(0, 0))
    
  })
  
  ##
  
  comparison_plot_exp_abs <- reactive({
    
    ggplot(data = prep_dat()) +
      geom_segment(data = prep_dat(), aes(x = Start, y = abs_frac_exch_state, xend = End, yend = abs_frac_exch_state, color = State)) +
      geom_errorbar(data = prep_dat(), aes(x = Med_Sequence, ymin = abs_frac_exch_state - err_abs_frac_exch_state, ymax = abs_frac_exch_state + err_abs_frac_exch_state, color = State)) +
      theme(legend.position = "bottom",
            legend.title = element_blank()) +
      scale_y_continuous(expand = c(0, 0))
    
  })
  
  ##
  
  cp_out <- reactive({
    
    comparison_plot_colors_chosen()
    
    if (input[["theory"]]) {
      
      if (input[["calc_type"]] == "relative") {
        
        cp <- comparison_plot_theo() 
        
      } else {
        
        cp <- comparison_plot_theo_abs() 
      }
      
    } else {
      
      if (input[["calc_type"]] == "relative") {
        
        cp <- comparison_plot_exp() 
        
      } else {
        
        cp <- comparison_plot_exp_abs()
        
      }
      
    }
    
    cp + coord_cartesian(xlim = c(input[["plot_x_range"]][[1]], input[["plot_x_range"]][[2]]),
                         ylim = c(input[["comp_plot_y_range"]][[1]], input[["comp_plot_y_range"]][[2]])) +
      labs(title = input[["comparison_plot_title"]], 
           x = input[["comparison_plot_x_label"]],
           y = input[["comparison_plot_y_label"]]) +
      scale_color_manual(values = comparison_plot_colors_chosen())
    
  })
  
  ##
  
  output[["comparisonPlot"]] <- renderPlot({

    cp_out() 
    
  })
  
  ##
  
  output[["comparisonPlot_debug"]] <- renderUI({
    
    if(!is.null(input[["comparisonPlot_hover"]])) {
      
      plot_data <- cp_out()[["data"]]
      hv <- input[["comparisonPlot_hover"]]
      
      hv_dat <- data.frame(x = hv[["x"]],
                           y = hv[["y"]],
                           Start = plot_data[[hv[["mapping"]][["x"]]]],
                           End = plot_data[["End"]],
                           y_plot = plot_data[[hv[["mapping"]][["y"]]]],
                           Sequence = plot_data[["Sequence"]],
                           State = plot_data[["State"]])
      
      tt_df <- filter(hv_dat, Start < x, End > x) %>% 
        filter(abs(y_plot - y) < 10) %>%
        filter(abs(y_plot - y) == min(abs(y_plot - y)))
   
      
      if(nrow(tt_df) != 0) { 
        
        tt_pos_adj <- ifelse(hv[["coords_img"]][["x"]]/hv[["range"]][["right"]] < 0.5,
                             "left", "right")
        
        tt_pos <- ifelse(hv[["coords_img"]][["x"]]/hv[["range"]][["right"]] < 0.5,
                         hv[["coords_css"]][["x"]], 
                         hv[["range"]][["right"]]/hv[["img_css_ratio"]][["x"]] - hv[["coords_css"]][["x"]])
        
        
        style <- paste0("position:absolute; z-index:1000; background-color: rgba(245, 245, 245, 1); ",
                        tt_pos_adj, ":", tt_pos, 
                        "px; top:", hv[["coords_css"]][["y"]], "px; padding: 0px;")
        
        div(
          style = style,
          p(HTML(paste0(tt_df[["Sequence"]], 
                        "<br/> Position: ", tt_df[["Start"]], "-", tt_df[["End"]], 
                        "<br/> Value: ", round(tt_df[["y_plot"]], 2),
                        "<br/> State: ", tt_df[["State"]])))
        )
      }
    }
  })
  
  ##
  
  output[["comparisonPlot_download_button"]] <- downloadHandler("comparisonPlot.svg",
                                                                content = function(file) {
                                                                  ggsave(file, cp_out(), device = svg,
                                                                         height = 300, width = 400, units = "mm")
                                                                })
  
  ##
  
  comparison_plot_data_theo <- reactive({
    
    prep_dat() %>%
      select(Protein, Sequence, State, Start, End, avg_theo_in_time, err_avg_theo_in_time) %>%
      filter(Protein == input[["chosen_protein"]],
             Start >= input[["plot_x_range"]][[1]],
             End <= input[["plot_x_range"]][[2]]) %>%
      mutate(avg_theo_in_time = round(avg_theo_in_time, 4),
             err_avg_theo_in_time = round(err_avg_theo_in_time, 4)) %>%
      arrange(Start, End) %>%
      dt_format(cols = c("Protein", "Sequence", "State", "Start", "End", "Theo Frac Exch", "Err Theo Frac Exch"))
    
  })
  
  ##
  
  comparison_plot_data_theo_abs <- reactive({
    
    prep_dat() %>%
      select(Protein, Sequence, State, Start, End, abs_avg_theo_in_time, err_abs_avg_theo_in_time) %>%
      filter(Protein == input[["chosen_protein"]],
             Start >= input[["plot_x_range"]][[1]],
             End <= input[["plot_x_range"]][[2]]) %>%
      mutate(abs_avg_theo_in_time = round(abs_avg_theo_in_time, 4),
             err_abs_avg_theo_in_time = round(err_abs_avg_theo_in_time, 4)) %>%
      arrange(Start, End) %>%
      dt_format(cols = c("Protein", "Sequence", "State", "Start", "End", "Theo Abs Val Exch", "Err Theo Abs Val Exch"))
  })
  
  ## 
  
  comparison_plot_data_exp <- reactive({
    
    prep_dat() %>%
      select(Protein, Sequence, State, Start, End, frac_exch_state, err_frac_exch_state) %>%
      filter(Protein == input[["chosen_protein"]],
             Start >= input[["plot_x_range"]][[1]],
             End <= input[["plot_x_range"]][[2]]) %>%
      mutate(frac_exch_state = round(frac_exch_state, 4),
             err_frac_exch_state = round(err_frac_exch_state, 4)) %>%
      arrange(Start, End) %>%
      dt_format(cols = c("Protein", "Sequence", "State", "Start", "End", "Frac Exch", "Err Frac Exch"))
    
  })
  
  ##
  
  comparison_plot_data_exp_abs <- reactive({
    
    prep_dat() %>%
      select(Protein, Sequence, State, Start, End, abs_frac_exch_state, err_abs_frac_exch_state) %>%
      filter(Protein == input[["chosen_protein"]],
             Start >= input[["plot_x_range"]][[1]],
             End <= input[["plot_x_range"]][[2]]) %>%
      mutate(abs_frac_exch_state = round(abs_frac_exch_state, 4),
             err_abs_frac_exch_state = round(err_abs_frac_exch_state, 4)) %>%
      arrange(Start, End) %>%
      dt_format(cols = c("Protein", "Sequence", "State", "Start", "End", "Abs Val Exch", "Err Abs Val Exch"))
    
  })
  
  ##
  
  output[["comparisonPlot_data"]] <- DT::renderDataTable(server = FALSE, {
    
    if (input[["theory"]]) {
      
      if (input[["calc_type"]] == "relative") {
        comparison_plot_data_theo()  
      } else {
        comparison_plot_data_theo_abs()
      }
      
    } else {
      
      if (input[["calc_type"]] == "absolute") {
        comparison_plot_data_exp_abs()
      } else {
        comparison_plot_data_exp()
      }
      
    }
    
  })
  
  ##
  
  ## WOODS PLOT + DATA
  
  ##
  
  woods_plot_dat <- reactive({
    
    validate(need(input[["compare_states"]], "Please select at least one state."))
    validate(need(length(unique(filter(dat(), !is.na("Modification"), Protein == input[["chosen_protein"]])[["State"]])) > 1, "Not sufficient number of states without modifications."))
    
    if(!input[["theory"]]){
      validate(need(as.numeric(input[["in_time"]]) < as.numeric(input[["chosen_time"]]), "In time must be smaller than chosen time."))
      validate(need(as.numeric(input[["chosen_time"]]) < as.numeric(input[["out_time"]]), "Out time must be bigger than chosen time."))
    }
    
    
    tmp <- bind_rows(lapply(c(input[["state_first"]], input[["state_second"]]), function(i) calculate_state_deuteration(dat(), 
                                                                                                                        protein = input[["chosen_protein"]], 
                                                                                                                        state = i, 
                                                                                                                        time_in = input[["in_time"]],
                                                                                                                        time_chosen = input[["chosen_time"]], 
                                                                                                                        time_out = input[["out_time"]],
                                                                                                                        deut_part = 0.01*as.integer(input[["deut_concentration"]])))) %>%
      droplevels() %>% 
      mutate(State = factor(State, levels = c(input[["state_first"]], input[["state_second"]]), labels = c("1", "2"))) %>%
      gather(variable, value, -c(Protein:End, State, Med_Sequence)) %>%
      unite(tmp, variable, State) %>%
      spread(tmp, value) 
    
    validate(need(!is.na(tmp[["frac_exch_state_1"]]), "First state data is not sufficient. Choose another state."))
    validate(need(!is.na(tmp[["frac_exch_state_2"]]), "Second state data is not sufficient. Choose another state."))
    
    tmp %>%
      mutate(diff_frac_exch = frac_exch_state_1 - frac_exch_state_2,
             err_frac_exch = sqrt(err_frac_exch_state_1^2 + err_frac_exch_state_2^2),
             abs_diff_frac_exch = abs_frac_exch_state_1 - abs_frac_exch_state_2,
             err_abs_diff_frac_exch = sqrt(err_abs_frac_exch_state_1^2 + err_abs_frac_exch_state_2^2),
             diff_theo_frac_exch = avg_theo_in_time_1 - avg_theo_in_time_2, 
             err_diff_theo_frac_exch = sqrt(err_avg_theo_in_time_1^2 + err_avg_theo_in_time_2^2),
             abs_diff_theo_frac_exch = abs_avg_theo_in_time_1 - abs_avg_theo_in_time_2,
             err_abs_diff_theo_frac_exch = sqrt(err_abs_avg_theo_in_time_1^2 + err_abs_avg_theo_in_time_2^2)) %>%
      select(Protein, Start, End, Med_Sequence, everything(), -contains("1"), -contains("2"))
    
  })
  
  ##
  
  differential_plot_theo <- reactive({
    
    confidence_limit <- as.double(input[["confidence_limit"]])
    confidence_limit_2 <- as.double(input[["confidence_limit_2"]])
    
    interval <- calculate_confidence_limit_values(calc_dat = woods_plot_dat(),
                                                  confidence_limit = confidence_limit,
                                                  theoretical = TRUE,
                                                  relative = TRUE)
    
    interval_2 <- calculate_confidence_limit_values(calc_dat = woods_plot_dat(),
                                                    confidence_limit = confidence_limit_2,
                                                    theoretical = TRUE,
                                                    relative = TRUE)
    
    mutate(woods_plot_dat(), colour = case_when(
      woods_plot_dat()[["diff_theo_frac_exch"]] < interval_2[1] ~ "deepskyblue3",
      woods_plot_dat()[["diff_theo_frac_exch"]] < interval[1] ~ "deepskyblue1",
      woods_plot_dat()[["diff_theo_frac_exch"]] > interval_2[2] ~ "firebrick3",
      woods_plot_dat()[["diff_theo_frac_exch"]] > interval[2] ~ "firebrick1",
      TRUE ~ "azure3")) %>%
      ggplot() +
      geom_segment(aes(x = Start, y = diff_theo_frac_exch, xend = End, yend = diff_theo_frac_exch, color = colour)) +
      geom_errorbar(aes(x = Med_Sequence, ymin = diff_theo_frac_exch - err_diff_theo_frac_exch, ymax = diff_theo_frac_exch + err_diff_theo_frac_exch, color = colour)) +
      geom_hline(yintercept = 0, linetype = "dotted", color = "green", size = .7) +
      geom_hline(aes(yintercept = interval[1], linetype = paste0(" Confidence interval ", confidence_limit*100, "% : ", round(interval[2], 4))), color = "deepskyblue1", size = .7, show.legend = TRUE) + 
      geom_hline(aes(yintercept = interval[2], linetype = paste0(" Confidence interval ", confidence_limit*100, "% : ", round(interval[2], 4))), color = "firebrick1", size = .7, show.legend = FALSE) +
      geom_hline(aes(yintercept = interval_2[1], linetype = paste0(" Confidence interval ", confidence_limit_2*100, "% : ", round(interval_2[2], 4))), color = "deepskyblue3", size = .7, show.legend = TRUE) +
      geom_hline(aes(yintercept = interval_2[2], linetype = paste0(" Confidence interval ", confidence_limit_2*100, "% : ", round(interval_2[2], 4))), color = "firebrick3", size = .7, show.legend = FALSE) +
      scale_linetype_manual(values = c("dashed", "dotdash")) + 
      scale_colour_identity() +
      scale_y_continuous(expand = c(0, 0), limits = c(-100, 100)) +
      theme(legend.title = element_blank(),
            legend.position = "bottom",
            legend.direction = "vertical") 
  })
  
  ##
  
  differential_plot_theo_abs <- reactive({
    
    confidence_limit <- as.double(input[["confidence_limit"]])
    confidence_limit_2 <- as.double(input[["confidence_limit_2"]])
    
    interval <- calculate_confidence_limit_values(calc_dat = woods_plot_dat(),
                                                  confidence_limit = confidence_limit,
                                                  theoretical = TRUE,
                                                  relative = FALSE)
    
    interval_2 <- calculate_confidence_limit_values(calc_dat = woods_plot_dat(),
                                                    confidence_limit = confidence_limit_2,
                                                    theoretical = TRUE,
                                                    relative = FALSE)
    
    mutate(woods_plot_dat(), colour = case_when(
      woods_plot_dat()[["abs_diff_theo_frac_exch"]] < interval_2[1] ~ "deepskyblue3",
      woods_plot_dat()[["abs_diff_theo_frac_exch"]] < interval[1] ~ "deepskyblue1",
      woods_plot_dat()[["abs_diff_theo_frac_exch"]] > interval_2[2] ~ "firebrick3",
      woods_plot_dat()[["abs_diff_theo_frac_exch"]] > interval[2] ~ "firebrick1",
      TRUE ~ "azure3")) %>%
      ggplot() +
      geom_segment(aes(x = Start, y = abs_diff_theo_frac_exch, xend = End, yend = abs_diff_theo_frac_exch, color = colour)) +
      geom_errorbar(aes(x = Med_Sequence, ymin = abs_diff_theo_frac_exch - err_abs_diff_theo_frac_exch, ymax = abs_diff_theo_frac_exch + err_abs_diff_theo_frac_exch, color = colour)) +
      geom_hline(yintercept = 0, linetype = "dotted", color = "green", size = .7) +
      geom_hline(aes(yintercept = interval[1], linetype = paste0(" Confidence interval ", confidence_limit*100, "% : ", round(interval[2], 4))), color = "deepskyblue1", size = .7, show.legend = TRUE) + 
      geom_hline(aes(yintercept = interval[2], linetype = paste0(" Confidence interval ", confidence_limit*100, "% : ", round(interval[2], 4))), color = "firebrick1", size = .7, show.legend = FALSE) +
      geom_hline(aes(yintercept = interval_2[1], linetype = paste0(" Confidence interval ", confidence_limit_2*100, "% : ", round(interval_2[2], 4))), color = "deepskyblue3", size = .7, show.legend = TRUE) +
      geom_hline(aes(yintercept = interval_2[2], linetype = paste0(" Confidence interval ", confidence_limit_2*100, "% : ", round(interval_2[2], 4))), color = "firebrick3", size = .7, show.legend = FALSE) +
      scale_linetype_manual(values = c("dashed", "dotdash")) + 
      scale_colour_identity() +
      scale_y_continuous(expand = c(0, 0), limits = c(-1, 1)) +
      theme(legend.title = element_blank(),
            legend.position = "bottom",
            legend.direction = "vertical") 
  })
  
  ##
  
  differential_plot_exp <- reactive({
    
    confidence_limit <- as.double(input[["confidence_limit"]])
    confidence_limit_2 <- as.double(input[["confidence_limit_2"]])
    
    interval <- calculate_confidence_limit_values(calc_dat = woods_plot_dat(),
                                                  confidence_limit = confidence_limit,
                                                  theoretical = FALSE,
                                                  relative = TRUE)
    
    interval_2 <- calculate_confidence_limit_values(calc_dat = woods_plot_dat(),
                                                    confidence_limit = confidence_limit_2,
                                                    theoretical = FALSE,
                                                    relative = TRUE)
    
    mutate(woods_plot_dat(), colour = case_when(
      woods_plot_dat()[["diff_frac_exch"]] < interval_2[1] ~ "deepskyblue3",
      woods_plot_dat()[["diff_frac_exch"]] < interval[1] ~ "deepskyblue1",
      woods_plot_dat()[["diff_frac_exch"]] > interval_2[2] ~ "firebrick3",
      woods_plot_dat()[["diff_frac_exch"]] > interval[2] ~ "firebrick1",
      TRUE ~ "azure3")) %>%
      ggplot() +
      geom_segment(aes(x = Start, y = diff_frac_exch, xend = End, yend = diff_frac_exch, color = colour)) +
      geom_errorbar(aes(x = Med_Sequence, ymin = diff_frac_exch - err_frac_exch, ymax = diff_frac_exch + err_frac_exch, color = colour)) +
      geom_hline(yintercept = 0, linetype = "dotted", color = "green", size = .7) +
      geom_hline(aes(yintercept = interval[1], linetype = paste0(" Confidence interval ", confidence_limit*100, "% : ", round(interval[2], 4))), color = "deepskyblue1", size = .7, show.legend = TRUE) +
      geom_hline(aes(yintercept = interval[2], linetype = paste0(" Confidence interval ", confidence_limit*100, "% : ", round(interval[2], 4))), color = "firebrick1", size = .7, show.legend = FALSE) +
      geom_hline(aes(yintercept = interval_2[1], linetype = paste0(" Confidence interval ", confidence_limit_2*100, "% : ", round(interval_2[2], 4))), color = "deepskyblue3", size = .7, show.legend = TRUE) +
      geom_hline(aes(yintercept = interval_2[2], linetype = paste0(" Confidence interval ", confidence_limit_2*100, "% : ", round(interval_2[2], 4))), color = "firebrick3", size = .7, show.legend = FALSE) +
      scale_linetype_manual(values = c("dashed", "dotdash")) +
      scale_colour_identity() +
      scale_y_continuous(expand = c(0, 0), limits = c(-100, 100)) +
      theme(legend.title = element_blank(),
            legend.position = "bottom",
            legend.direction = "vertical") 
  })
  
  ##
  
  differential_plot_exp_abs <- reactive({
    
    confidence_limit <- as.double(input[["confidence_limit"]])
    confidence_limit_2 <- as.double(input[["confidence_limit_2"]])
    
    interval <- calculate_confidence_limit_values(calc_dat = woods_plot_dat(),
                                                  confidence_limit = confidence_limit,
                                                  theoretical = FALSE,
                                                  relative = FALSE)
    
    interval_2 <- calculate_confidence_limit_values(calc_dat = woods_plot_dat(),
                                                    confidence_limit = confidence_limit_2,
                                                    theoretical = FALSE,
                                                    relative = FALSE)
    
    mutate(woods_plot_dat(), colour = case_when(
      woods_plot_dat()[["abs_diff_frac_exch"]] < interval_2[1] ~ "deepskyblue3",
      woods_plot_dat()[["abs_diff_frac_exch"]] < interval[1] ~ "deepskyblue1",
      woods_plot_dat()[["abs_diff_frac_exch"]] > interval_2[2] ~ "firebrick3",
      woods_plot_dat()[["abs_diff_frac_exch"]] > interval[2] ~ "firebrick1",
      TRUE ~ "azure3")) %>%
      ggplot() +
      geom_segment(aes(x = Start, y = abs_diff_frac_exch, xend = End, yend = abs_diff_frac_exch, color = colour)) +
      geom_errorbar(aes(x = Med_Sequence, ymin = abs_diff_frac_exch - err_abs_diff_frac_exch, ymax = abs_diff_frac_exch + err_abs_diff_frac_exch, color = colour)) +
      geom_hline(yintercept = 0, linetype = "dotted", color = "green", size = .7) +
      geom_hline(aes(yintercept = interval[1], linetype = paste0(" Confidence interval ", confidence_limit*100, "% : ", round(interval[2], 4))), color = "deepskyblue1", size = .7, show.legend = TRUE) + 
      geom_hline(aes(yintercept = interval[2], linetype = paste0(" Confidence interval ", confidence_limit*100, "% : ", round(interval[2], 4))), color = "firebrick1", size = .7, show.legend = FALSE) +
      geom_hline(aes(yintercept = interval_2[1], linetype = paste0(" Confidence interval ", confidence_limit_2*100, "% : ", round(interval_2[2], 4))), color = "deepskyblue3", size = .7, show.legend = TRUE) +
      geom_hline(aes(yintercept = interval_2[2], linetype = paste0(" Confidence interval ", confidence_limit_2*100, "% : ", round(interval_2[2], 4))), color = "firebrick3", size = .7, show.legend = FALSE) +
      scale_linetype_manual(values = c("dashed", "dotdash")) + 
      scale_colour_identity() +
      scale_y_continuous(expand = c(0, 0), limits = c(-1, 1)) +
      theme(legend.title = element_blank(),
            legend.position = "bottom",
            legend.direction = "vertical") 
  })
  
  ##
  
  wp_out <- reactive({
    
    validate(need(!(input[["state_first"]] == input[["state_second"]]), "Please select two different states."))
    # validate(need(differential_plot_theo(), "Select different states."))
    
    if (input[["theory"]]) {
      
      if (input[["calc_type"]] == "relative") {
        
        wp <- differential_plot_theo() 
        
      } else {
        
        wp <- differential_plot_theo_abs() 
        
      }
      
    } else {
      
      if (input[["calc_type"]] == "relative") {
        
        wp <- differential_plot_exp() 
        
      } else {
        
        wp <- differential_plot_exp_abs() 
        
      }
      
    }
    
    wp + coord_cartesian(xlim = c(input[["plot_x_range"]][[1]], input[["plot_x_range"]][[2]]),
                         ylim = c(input[["woods_plot_y_range"]][[1]], input[["woods_plot_y_range"]][[2]])) +
      labs(title = input[["woods_plot_title"]],
           x = input[["woods_plot_x_label"]],
           y = input[["woods_plot_y_label"]])
    
  })
  
  ##
  
  output[["differentialPlot"]] <- renderPlot({
    
    wp_out()
    
  })
  
  ##
  
  output[["differentialPlot_debug"]] <- renderUI({
    
    if(!is.null(input[["differentialPlot_hover"]])) {
      
      wp_plot_data <- wp_out()[["data"]]
      wp_hv <- input[["differentialPlot_hover"]]

      wp_hv_dat <- data.frame(x = wp_hv[["x"]],
                           y = wp_hv[["y"]],
                           Start = wp_plot_data[[wp_hv[["mapping"]][["x"]]]],
                           End = wp_plot_data[["End"]],
                           y_plot = wp_plot_data[[wp_hv[["mapping"]][["y"]]]],
                           Sequence = wp_plot_data[["Sequence"]])
      
      wp_tt_df <- filter(wp_hv_dat, Start < x, End > x) %>% 
        filter(abs(y_plot - y) < 10) %>%
        filter(abs(y_plot - y) == min(abs(y_plot - y)))
      
      if(nrow(wp_tt_df) != 0) {

        wp_tt_pos_adj <- ifelse(wp_hv[["coords_img"]][["x"]]/wp_hv[["range"]][["right"]] < 0.5,
                             "left", "right")
        
        wp_tt_pos <- ifelse(wp_hv[["coords_img"]][["x"]]/wp_hv[["range"]][["right"]] < 0.5,
                            wp_hv[["coords_css"]][["x"]], 
                            wp_hv[["range"]][["right"]]/wp_hv[["img_css_ratio"]][["x"]] - wp_hv[["coords_css"]][["x"]])
        
       
        style <- paste0("position:absolute; z-index:1072; background-color: rgba(245, 245, 245, 1); pointer-events: none; ",
                        wp_tt_pos_adj, ":", wp_tt_pos, "px; padding: 0px;",
                        "top:", wp_hv[["coords_css"]][["y"]] , "px; ") 
        
        wellPanel(
          style = style,
          p(HTML(paste0(wp_tt_df[["Sequence"]],
                        "<br/> Position: ", wp_tt_df[["Start"]], "-", wp_tt_df[["End"]],
                        "<br/> Value: ", round(wp_tt_df[["y_plot"]], 2))))
        )
        
      }
    }
  })
  
  ##
  
  output[["differentialPlot_download_button"]] <- downloadHandler("differentialPlot.svg",
                                                                  content = function(file) {
                                                                    ggsave(file, wp_out(), device = svg,
                                                                           height = 300, width = 400, units = "mm")
                                                                  })
  
  ##
  
  differential_plot_data_theo <- reactive({
    
    woods_plot_dat() %>%
      add_stat_dependency(confidence_limit = as.double(input[["confidence_limit"]]),
                          theoretical = TRUE, 
                          relative = TRUE) %>%
      add_stat_dependency(confidence_limit = as.double(input[["confidence_limit_2"]]),
                          theoretical = TRUE, 
                          relative = TRUE) %>%
      select(Protein, Sequence, Start, End, diff_theo_frac_exch, err_diff_theo_frac_exch, paste0("valid_at_", input[["confidence_limit"]]), paste0("valid_at_", input[["confidence_limit_2"]])) %>%
      filter(Protein == input[["chosen_protein"]],
             Start >= input[["plot_x_range"]][[1]],
             End <= input[["plot_x_range"]][[2]]) %>%
      mutate(diff_theo_frac_exch = round(diff_theo_frac_exch, 4),
             err_diff_theo_frac_exch = round(err_diff_theo_frac_exch, 4)) %>%
      arrange(Start, End) %>%
      dt_format(cols = unique(c("Protein", "Sequence", "Start", "End", "Theo Diff Frac Exch", "Err Theo Diff Frac Exch", paste0("Valid At ", input[["confidence_limit"]]), paste0("Valid At ", input[["confidence_limit_2"]]))))
    
  })
  
  ##
  
  differential_plot_data_theo_abs <- reactive({
    
    woods_plot_dat() %>%
      add_stat_dependency(confidence_limit = as.double(input[["confidence_limit"]]),
                          theoretical = TRUE, 
                          relative = FALSE) %>%
      add_stat_dependency(confidence_limit = as.double(input[["confidence_limit_2"]]),
                          theoretical = TRUE, 
                          relative = FALSE) %>%
      select(Protein, Sequence, Start, End, abs_diff_theo_frac_exch, err_abs_diff_theo_frac_exch, paste0("valid_at_", input[["confidence_limit"]]), paste0("valid_at_", input[["confidence_limit_2"]])) %>%
      filter(Protein == input[["chosen_protein"]],
             Start >= input[["plot_x_range"]][[1]],
             End <= input[["plot_x_range"]][[2]]) %>%
      mutate(abs_diff_theo_frac_exch = round(abs_diff_theo_frac_exch, 4),
             err_abs_diff_theo_frac_exch = round(err_abs_diff_theo_frac_exch, 4)) %>%
      arrange(Start, End) %>%
      dt_format(cols = unique(c("Protein", "Sequence", "Start", "End", "Theo Abs Value Diff", "Err Theo Abs Value Diff", paste0("Valid At ", input[["confidence_limit"]]), paste0("Valid At ", input[["confidence_limit_2"]]))))
    
  })
  
  ##
  
  differential_plot_data_exp <- reactive({
    
    woods_plot_dat() %>%
      add_stat_dependency(confidence_limit = as.double(input[["confidence_limit"]]),
                          theoretical = FALSE, 
                          relative = TRUE) %>%
      add_stat_dependency(confidence_limit = as.double(input[["confidence_limit_2"]]),
                          theoretical = FALSE, 
                          relative = TRUE) %>%
      select(Protein, Sequence, Start, End, diff_frac_exch, err_frac_exch, paste0("valid_at_", input[["confidence_limit"]]), paste0("valid_at_", input[["confidence_limit_2"]])) %>%
      filter(Protein == input[["chosen_protein"]],
             Start >= input[["plot_x_range"]][[1]],
             End <= input[["plot_x_range"]][[2]]) %>%
      mutate(diff_frac_exch = round(diff_frac_exch, 4),
             err_frac_exch = round(err_frac_exch, 4)) %>%
      arrange(Start, End) %>%
      dt_format(cols = unique(c("Protein", "Sequence", "Start", "End", "Diff Frac Exch", "Err Diff Frac Exch", paste0("Valid At ", input[["confidence_limit"]]), paste0("Valid At ", input[["confidence_limit_2"]]))))
    
  })
  
  ##
  
  differential_plot_data_exp_abs <- reactive({
    
    woods_plot_dat() %>%
      add_stat_dependency(confidence_limit = as.double(input[["confidence_limit"]]),
                          theoretical = FALSE,
                          relative = FALSE) %>%
      add_stat_dependency(confidence_limit = as.double(input[["confidence_limit_2"]]),
                          theoretical = FALSE, 
                          relative = FALSE) %>%
      select(Protein, Sequence, Start, End, abs_diff_frac_exch, err_abs_diff_frac_exch, paste0("valid_at_", input[["confidence_limit"]]), paste0("valid_at_", input[["confidence_limit_2"]])) %>%
      filter(Protein == input[["chosen_protein"]],
             Start >= input[["plot_x_range"]][[1]],
             End <= input[["plot_x_range"]][[2]]) %>%
      mutate(abs_diff_frac_exch = round(abs_diff_frac_exch, 4),
             err_abs_diff_frac_exch = round(err_abs_diff_frac_exch, 4)) %>%
      arrange(Start, End) %>%
      dt_format(cols = unique(c("Protein", "Sequence", "Start", "End", "Diff Abs Value Exch", "Err Diff Abs Value Exch", paste0("Valid At ", input[["confidence_limit"]]), paste0("Valid At ", input[["confidence_limit_2"]]))))
  })
  
  ##
  
  output[["differentialPlot_data"]] <- DT::renderDataTable(server = FALSE, {
    
    if (input[["theory"]]) {
      
      if(input[["calc_type"]] == "relative") {
        differential_plot_data_theo()  
      } else {
        differential_plot_data_theo_abs()
      }
      
    } else {
      
      if (input[["calc_type"]] == "relative") {
        differential_plot_data_exp()  
      } else {
        differential_plot_data_exp_abs()
      }
      
    }
    
  })
  
  ##
  
  ### TAB: KINETICS ###
  
  ##
  
  observe({
    
    times_from_file <- unique(round(dat()[["Exposure"]], 3))
    
    tmp <- unique(round(dat()[["Exposure"]], 3))
    choose_time_out <- setNames(tmp, c(head(tmp, -1), "chosen control"))
    
    updateSelectInput(session, 
                      inputId = "kin_in_time",
                      choices = times_from_file[times_from_file < 99999],
                      selected = min(times_from_file[times_from_file > 0]))
    
    if(!has_modifications()){
      
      updateSelectInput(session, 
                        inputId = "kin_out_time",
                        choices =  choose_time_out,
                        selected = choose_time_out["chosen control"])
    }
    
    if(has_modifications()){
      
      updateSelectInput(session, 
                        inputId = "kin_out_time",
                        choices =  times_from_file[times_from_file < 99999],
                        selected = max(times_from_file[times_from_file < 99999]))
    }
    
    
  })
  
  ##
  
  observe({
    
    updateTextInput(session, 
                    inputId = "kin_plot_title",
                    value = case_when(
                      input[["kin_theory"]] ~ paste0("Theoretical uptake curve for chosen peptides for ", input[["chosen_protein"]]),
                      !input[["kin_theory"]]  ~ paste0("Uptake curve for chosen peptides for ", input[["chosen_protein"]])
                    ))
    
    updateTextInput(session, 
                    inputId = "kin_plot_y_label", 
                    value = case_when(
                      input[["kin_theory"]] & input[["kin_calc_type"]] == "relative" ~ "Theoretical deuteration [%]",
                      input[["kin_theory"]] & input[["kin_calc_type"]] == "absolute" ~ "Theoretical deuteration [Da]",
                      !input[["kin_theory"]] & input[["kin_calc_type"]] == "relative" ~ "Deuteration [%]",
                      !input[["kin_theory"]] & input[["kin_calc_type"]] == "absolute" ~ "Deuteration [Da]"
                    ))
    
  })
  
  ##
  
  observe({
    
    if(input[["kin_theory"]]){
      hide(id = "kin_time_part")
    }
    
  })
  
  observe({
    
    if(!input[["kin_theory"]]){
      show(id = "kin_time_part")
    }
    
  })
  
  ##
  
  observe({
    
    if(input[["kin_calc_type"]] == "absolute"){
      hide(id = "kin_out_time_part")
    } 
    
  })
  
  ##
  
  observe({
    
    if(input[["kin_calc_type"]] == "relative"){
      show(id = "kin_out_time_part")
    }
    
  })
  
  ##
  
  peptide_list <- reactive({
    
    dat() %>%
      filter(Protein == input[["chosen_protein"]]) %>%
      select(Sequence, State, Start, End) %>%
      unique(.) %>%
      arrange(Start, End)
    
  })
  
  ##
  
  output[["peptide_list_data"]] <- DT::renderDataTable({
    
    datatable(data = peptide_list(),
              class = "table-bordered table-condensed",
              extensions = "Buttons",
              options = list(pageLength = 10, dom = "tip", autoWidth = TRUE, target = 'cell'),
              filter = "bottom",
              rownames = FALSE)
    
  })
  
  ##
  
  DTproxy <- DT::dataTableProxy("peptide_list_data", session = session)
  
  ##
  
  observeEvent(input[["reset_peptide_list"]], {
    
    DT::selectRows(DTproxy, NULL)
    
  })
  
  ##
  
  kin_dat <- reactive({
    
    validate(need(input[["peptide_list_data_rows_selected"]], "Please select at least one peptide from the table on the left."))
    
    times_from_file <- unique(round(dat()[["Exposure"]], 3))
    
    if(input[["kin_theory"]]){
 
      bind_rows(apply(peptide_list()[input[["peptide_list_data_rows_selected"]], ], 1, function(peptide){
        calculate_kinetics(dat = dat(),
                           protein = input[["chosen_protein"]], 
                           sequence = peptide[1],
                           state = peptide[2],
                           start = as.numeric(peptide[3]),
                           end = as.numeric(peptide[4]),
                           time_in = min(times_from_file[times_from_file > 0]),
                           time_out = max(times_from_file),
                           deut_part = 0.01*as.integer(input[["deut_concentration"]]))
      }))
      
    } else {
      
      validate(need(as.numeric(input[["kin_out_time"]]) > as.numeric(input[["kin_in_time"]]), "Out time must be bigger than in time. "))
      
      validate(need(sum(times_from_file < as.numeric(input[["kin_out_time"]]) & times_from_file > as.numeric(input[["kin_in_time"]])) > 1, "Not enough time points between in and out time. "))

      bind_rows(apply(peptide_list()[input[["peptide_list_data_rows_selected"]], ], 1, function(peptide){
        calculate_kinetics(dat = dat(),
                           protein = input[["chosen_protein"]], 
                           sequence = peptide[1],
                           state = peptide[2],
                           start = as.numeric(peptide[3]),
                           end = as.numeric(peptide[4]),
                           time_in = as.numeric(input[["kin_in_time"]]),
                           time_out = as.numeric(input[["kin_out_time"]]),
                           deut_part = 0.01*as.integer(input[["deut_concentration"]]))
      }))
      
    }
  
    
    
    
  })
  
  ##
  
  observe({
    
    if (input[["kin_calc_type"]] == "absolute") {
      
      min_kin_abs <- round_any(min(kin_dat()[c("abs_frac_exch_state", "abs_avg_theo_in_time")], na.rm = TRUE), 5, floor)
      max_kin_abs <- round_any(max(kin_dat()[c("abs_frac_exch_state", "abs_avg_theo_in_time")], na.rm = TRUE), 5, ceiling)
      
      updateSliderInput(session,
                        inputId = "kin_plot_y_range",
                        min = 0,
                        max = max_kin_abs + 5,
                        value = c(0, max_kin_abs),
                        step = 1)
      
    } else {
      
      updateSliderInput(session,
                        inputId = "kin_plot_y_range",
                        min = -50,
                        max = 200,
                        value = c(-10, 100),
                        step = 10)
    }
    
  })
  
  ##
  
  ##  KINETIC PLOT +  DATA
  
  ##
  
  kin_plot_theo <- reactive({
    
    kin_dat() %>% 
      mutate(prop = paste0(Sequence, "-", State)) %>%
      ggplot(aes(x = time_chosen, y = avg_theo_in_time, group = prop)) +
      geom_point() + 
      geom_ribbon(aes(ymin = avg_theo_in_time - err_avg_theo_in_time, ymax = avg_theo_in_time + err_avg_theo_in_time, fill = prop), alpha = 0.15) +
      geom_line(aes(color = prop))
    
  })
  
  ##
  
  kin_plot_theo_abs <- reactive({
    
    kin_dat() %>% 
      mutate(prop = paste0(Sequence, "-", State)) %>%
      ggplot(aes(x = time_chosen, y = abs_avg_theo_in_time, group = prop)) +
      geom_point() + 
      geom_ribbon(aes(ymin = abs_avg_theo_in_time - err_abs_avg_theo_in_time, ymax = abs_avg_theo_in_time + err_abs_avg_theo_in_time, fill = prop), alpha = 0.15) +
      geom_line(aes(color = prop))
    
  })
  
  ##
  
  kin_plot_exp <- reactive({
    
    
    kin_dat() %>% 
      mutate(prop = paste0(Sequence, "-", State)) %>%
      ggplot(aes(x = time_chosen, y = frac_exch_state, group = prop)) +
      geom_point() + 
      geom_ribbon(aes(ymin = frac_exch_state - err_frac_exch_state, ymax = frac_exch_state + err_frac_exch_state, fill = prop), alpha = 0.15) +
      geom_line(aes(color = prop))
    
  })
  
  ##
  
  kin_plot_exp_abs <- reactive({
    
    kin_dat() %>% 
      mutate(prop = paste0(Sequence, "-", State)) %>%
      ggplot(aes(x = time_chosen, y = abs_frac_exch_state, group = prop)) +
      geom_point() + 
      geom_ribbon(aes(ymin = abs_frac_exch_state - err_abs_frac_exch_state, ymax = abs_frac_exch_state + err_abs_frac_exch_state, fill = prop), alpha = 0.15) +
      geom_line(aes(color = prop))
    
  })
  
  ##
  
  kp_out <- reactive({
    
    if (input[["kin_theory"]]) {
      
      if (input[["kin_calc_type"]] == "relative") {
        
        kp <- kin_plot_theo()
        
      } else {
        
        kp <- kin_plot_theo_abs()
        
      }
      
    } else {
      
      if (input[["kin_calc_type"]] == "relative"){
        
        kp <- kin_plot_exp()
        
      } else {
        
        kp <- kin_plot_exp_abs()
        
      }
      
    }
    
    kp + 
      geom_point(size = 3) +
      labs(title = input[["kin_plot_title"]],
              x = input[["kin_plot_x_label"]],
              y = input[["kin_plot_y_label"]]) +
      coord_cartesian(ylim = c(input[["kin_plot_y_range"]][1], input[["kin_plot_y_range"]][2])) +
      scale_x_log10() + 
      theme(legend.position = "bottom",
            legend.title = element_blank())
    
  })
  
  ##
  
  output[["kinetic_plot_chosen_peptides"]] <- renderPlot({
    
    kp_out()
    
  })
  
  ##
  
  output[["kinetic_plot_chosen_peptides_debug"]] <- renderUI({
    
    if(!is.null(input[["kinetic_plot_chosen_peptides_hover"]])) {
      
      plot_data <- kp_out()[["data"]]
      hv <- input[["kinetic_plot_chosen_peptides_hover"]]
      
      hv_dat <- data.frame(x = hv[["x"]],
                           y = hv[["y"]],
                           Start = plot_data[["Start"]],
                           End = plot_data[["End"]],
                           x_plot = plot_data[[hv[["mapping"]][["x"]]]],
                           y_plot = plot_data[[hv[["mapping"]][["y"]]]],
                           Sequence = plot_data[["Sequence"]],
                           State = plot_data[["State"]])
      
      tt_df <- filter(hv_dat, abs(y_plot - y) < 10, abs(y_plot - y) == min(abs(y_plot - y))) %>%
                      filter(abs(x_plot - x) < 0.1*x_plot, abs(x_plot - x) == min(abs(x_plot - x))) 
      
      if(nrow(tt_df) != 0) { 
        
        tt_pos_adj <- ifelse(hv[["coords_img"]][["x"]]/hv[["range"]][["right"]] < 0.5,
                             "left", "right")
        
        tt_pos <- ifelse(hv[["coords_img"]][["x"]]/hv[["range"]][["right"]] < 0.5,
                         hv[["coords_css"]][["x"]], 
                         hv[["range"]][["right"]]/hv[["img_css_ratio"]][["x"]] - hv[["coords_css"]][["x"]])
        
        
        style <- paste0("position:absolute; z-index:1000; background-color: rgba(245, 245, 245, 1); pointer-events: none;",
                        tt_pos_adj, ":", tt_pos, 
                        "px; top:", hv[["coords_css"]][["y"]], "px; padding: 0px;")
        
        div(
          style = style,
          p(HTML(paste0(tt_df[["Sequence"]], 
                        "<br/> State: ", tt_df[["State"]],
                        "<br/> Position: ", tt_df[["Start"]], "-", tt_df[["End"]], 
                        "<br/> Value: ", round(tt_df[["y_plot"]], 2),
                        "<br/> Time point: ", tt_df[["x_plot"]], " min")))
        )
      }
    }
  })
  
  ##
  
  output[["kineticPlot_download_button"]] <- downloadHandler("kineticPlot.svg",
                                                             content = function(file){
                                                               ggsave(file, kp_out(), device = svg,
                                                                      height = 300, width = 400, units = "mm")
                                                             })
  ##
  
  kin_plot_exp_data <- reactive({
    
    kin_dat() %>%
      select(Protein, Sequence, State, Start, End, time_chosen, frac_exch_state, err_frac_exch_state) %>%
      mutate(frac_exch_state = round(frac_exch_state, 4), 
             err_frac_exch_state = round(err_frac_exch_state, 4)) %>%
      dt_format(cols = c("Protein", "Sequence", "State", "Start", "End", "Time Point", "Frac Exch", "Err Frac Exch"))
    
  })
  
  ##
  
  kin_plot_exp_abs_data <- reactive({
    
    kin_dat() %>%
      select(Protein, Sequence, State, Start, End, time_chosen, abs_frac_exch_state, err_abs_frac_exch_state) %>%
      mutate(abs_frac_exch_state = round(abs_frac_exch_state, 4), 
             err_abs_frac_exch_state = round(err_abs_frac_exch_state, 4)) %>%
      dt_format(cols = c("Protein", "Sequence", "State", "Start", "End", "Time Point", "Abs Val Exch", "Err Abs Val Exch"))
    
  })
  
  ## 
  
  kin_plot_theo_data <- reactive({
    
    kin_dat() %>%
      select(Protein, Sequence, State, Start, End, time_chosen, avg_theo_in_time, err_avg_theo_in_time) %>%
      mutate(avg_theo_in_time = round(avg_theo_in_time, 4), 
             err_avg_theo_in_time = round(err_avg_theo_in_time, 4)) %>%
      dt_format(cols = c("Protein", "Sequence", "State", "Start", "End", "Time Point", "Theo Frac Exch", "Theo Err Frac Exch"))
    
  })
  
  ##
  
  kin_plot_theo_abs_data <- reactive({
    
    kin_dat() %>%
      select(Protein, Sequence, State, Start, End, time_chosen, abs_avg_theo_in_time, err_abs_avg_theo_in_time) %>%
      mutate(abs_avg_theo_in_time = round(abs_avg_theo_in_time, 4), 
             err_abs_avg_theo_in_time = round(err_abs_avg_theo_in_time, 4)) %>%
      dt_format(cols = c("Protein", "Sequence", "State", "Start", "End", "Time Point", "Theo Abs Val Exch", "Theo Err Abs Val Exch"))
    
  })
  
  ##
  
  output[["kin_plot_data"]] <- DT::renderDataTable(server = FALSE, {
    
    if (input[["kin_theory"]]) {
      
      if (input[["kin_calc_type"]] == "relative") {
        kin_plot_theo_data()
      } else {
        kin_plot_theo_abs_data()
      } 
      
    } else {
      
      if (input[["kin_calc_type"]] == "relative") {
        kin_plot_exp_data()
      } else {
        kin_plot_exp_abs_data()
      } 
      
    }
    
  })
  
  ### TAB: QUALITY CONTROL  
  
  ##
  
  observe({
    
    times_from_file <- unique(round(dat()[["Exposure"]], 3))
    
    updateSelectInput(session, 
                      inputId = "qc_chosen_time",
                      choices = times_from_file[times_from_file < 99999],
                      selected = min(times_from_file[times_from_file >= 1]))
    
    updateSelectInput(session, 
                      inputId = "qc_in_time",
                      choices = times_from_file[times_from_file < 99999],
                      selected = min(times_from_file[times_from_file > 0]))
    
    updateSelectInput(session,
                      inputId = "qc_state_first",
                      choices = states_from_file(),
                      selected = states_from_file()[1])
    
    updateSelectInput(session,
                      inputId = "qc_state_second",
                      choices = states_from_file(),
                      selected = states_from_file()[length(states_from_file())])
    
  })
  
  ##
  
  quality_control_dat <- reactive({
    
    qc_dat <- dat() %>%
      filter(Exposure < 99999)
 
    validate(need(as.numeric(input[["qc_chosen_time"]]) > as.numeric(input[["qc_in_time"]]), "Chosen time must be bigger than in time. "))
    validate(need(sum(unique(qc_dat[["Exposure"]]) > as.numeric(input[["qc_chosen_time"]])) > 1, "Not enough time points (bigger than chosen time) to generate a plot. ")) 

    result <- quality_control(dat = qc_dat,
                              state_first = input[["qc_state_first"]],
                              state_second = input[["qc_state_second"]], 
                              chosen_time = as.numeric(input[["qc_chosen_time"]]), 
                              in_time = as.numeric(input[["qc_in_time"]])) %>%
      # to get the percentages in readable form
      mutate(avg_err_state_first = 100 * avg_err_state_first,
             sd_err_state_first = 100 * sd_err_state_first,
             avg_err_state_second = 100 * avg_err_state_second,
             sd_err_state_second = 100 * sd_err_state_second, 
             avg_diff = 100 * avg_diff, 
             sd_diff = 100 * sd_diff)
      
    
  })
  
  ##
  
  qc_out <- reactive({
    
    quality_control_dat() %>%
      gather(2:7, key = 'type', value = 'value') %>%
      filter(startsWith(type, "avg")) %>%
      ggplot(aes(x = out_time, y = value, group = type)) +
      geom_point(size = 3) +
      geom_line(aes(color = type)) +
      scale_colour_discrete(name = "Mean uncertainty of: ", labels = c("difference", "first state", "second state")) +
      scale_x_log10() + 
      labs(x = "Out time [min]",
           y = "Mean uncertainty [%]",
           title = "Quality control plot for experiment")
    
  })
  
  output[["quality_control_plot"]] <- renderPlot({
    
    qc_out()
    
  })
  
  ##
  
  quality_control_plot_data_out <- reactive({
    
    quality_control_dat() %>%
      select(out_time, avg_err_state_first, avg_err_state_second, avg_diff) %>%
      mutate(avg_err_state_first = round(avg_err_state_first, 2),
             avg_err_state_second = round(avg_err_state_second, 2),
             avg_diff = round(avg_diff, 2)) %>%
      dt_format(cols = c("Out time", "Mean error - first state [%]", "Mean error - second state [%]", "Mean error of difference [%]"))
    
  })
  
  ##
  
  output[["quality_control_plot_data"]] <- DT::renderDataTable({
    
    quality_control_plot_data_out()
    
  })
  
  ##
  
  output[["quality_control_plot_debug"]] <- renderUI({
    
    if(!is.null(input[["quality_control_plot_hover"]])) {
      
      plot_data <- qc_out()[["data"]]
      hv <- input[["quality_control_plot_hover"]]
      
      hv_dat <- data.frame(x = hv[["x"]],
                           y = hv[["y"]],
                           x_plot = plot_data[[hv[["mapping"]][["x"]]]],
                           y_plot = plot_data[[hv[["mapping"]][["y"]]]])
      
      tt_df <- hv_dat %>%
        filter(abs(y_plot - y) == min(abs(y_plot - y)), abs(x_plot - x) < 0.1*x_plot) 
      
      if(nrow(tt_df) != 0) { 
        
        tt_pos_adj <- ifelse(hv[["coords_img"]][["x"]]/hv[["range"]][["right"]] < 0.5,
                             "left", "right")
        
        tt_pos <- ifelse(hv[["coords_img"]][["x"]]/hv[["range"]][["right"]] < 0.5,
                         hv[["coords_css"]][["x"]], 
                         hv[["range"]][["right"]]/hv[["img_css_ratio"]][["x"]] - hv[["coords_css"]][["x"]])
        
        
        style <- paste0("position:absolute; z-index:1000; background-color: rgba(245, 245, 245, 1); ",
                        tt_pos_adj, ":", tt_pos, 
                        "px; top:", hv[["coords_css"]][["y"]], "px; padding: 0px;")
        
        div(
          style = style,
          p(HTML(paste0("<br/> x: ", round(tt_df[["x_plot"]], 0), " [min]",
                        "<br/> y: ", round(tt_df[["y_plot"]], 2), " [%] ")))
        )
      }
    }
  })
  
  ##
  
  ##
  
  output[["quality_control_plot_download_button"]] <- downloadHandler("qualityControlPlot.svg",
                                                                      content = function(file){
                                                                        ggsave(file, qc_out(), device = svg,
                                                                               height = 300, width = 400, units = "mm")
                                                                      })
  
  ### TAB: SUMMARY
  
  summary_data <- reactive({
    
    n_reps <- group_by(dat(), Protein, Start, End, Sequence, Modification, State, Exposure) %>%
      summarise(n_rep = length(unique(File))) %>%
      ungroup() %>% 
      pull(n_rep) %>% 
      table() %>% 
      sort(decreasing = TRUE) %>% 
      names() %>% 
      as.numeric()
    
    data.frame(Name = c("HDX time course", 
                        "Number of peptides",
                        "Sequence coverage",
                        "Average peptide length",
                        "Redundancy",
                        "Replicates",
                        #"Average standard deviation",
                        "Significant differences in HDX"), 
               Value = c(length(unique(dat()[["Exposure"]])) - 1, # we add control as an additional timepoint 
                         length(unique(dat()[["Sequence"]])), 
                         paste0(100*round(mean(stateOverlapDist_data()[["coverage"]] > 0), 4), "%"), 
                         round(mean(nchar(unique(dat()[["Sequence"]]))), 4), 
                         round(mean(stateOverlapDist_data()[["coverage"]]), 4), 
                         n_reps[1], 
                         #NA, 
                         paste0(input[["confidence_limit"]], " | ", input[["confidence_limit_2"]]))
    )
    
  })
  
  
  output[["summary_table"]] <- DT::renderDataTable(server = FALSE, {
    
    datatable(data = summary_data(),
              class = "table-bordered table-condensed",
              extensions = "Buttons",
              options = list(pageLength = 10, dom = "tBi", autoWidth = TRUE, buttons = c("excel", "pdf")),
              filter = "bottom",
              rownames = FALSE)
    
  })
  
  ### TAB: REPORT ###
  
  ##
  
  output[["export_action"]] <- downloadHandler(
    
    filename <- "HaDeX_Report.html",
    
    content <- function(file) {
      
      rmarkdown::render(input = "report_template.Rmd", 
                        output_file = file, quiet = TRUE)
      
    })
  
  ##
  
}