R/runGUI.R
In eMetaboHUB/MS-CleanR: Cleans MS Data for Easier Analyses
Defines functions
runGUI
Documented in
runGUI
#' Launch the Graphical User Interface (GUI)
#'
#' @export
runGUI <- function() {
# datasets loading
utils::data("mass_neutral_loss_neg", package="mscleanr", envir=environment())
mass_neutral_loss_neg <- get("mass_neutral_loss_neg", envir=environment()) # redundant, but makes the syntax checker happy
utils::data("mass_neutral_loss_pos", package="mscleanr", envir=environment())
mass_neutral_loss_pos <- get("mass_neutral_loss_pos", envir=environment()) # redundant, but makes the syntax checker happy
utils::data("mass_adducts_neg", package="mscleanr", envir=environment())
mass_adducts_neg <- get("mass_adducts_neg", envir=environment()) # redundant, but makes the syntax checker happy
utils::data("mass_adducts_pos", package="mscleanr", envir=environment())
mass_adducts_pos <- get("mass_adducts_pos", envir=environment()) # redundant, but makes the syntax checker happy
utils::data("mass_isotopes", package="mscleanr", envir=environment())
mass_isotopes <- get("mass_isotopes", envir=environment()) # redundant, but makes the syntax checker happy
# personalized elements
referenceInput <- function(id, text) {
shiny::tabPanel(text,
shiny::fluidRow(
shiny::column(8,
shiny::fileInput(id,
paste0("(OPTIONAL) Upload a CSV reference file for ", tolower(text), "."),
accept = ".csv",
placeholder = " No file selected",
width = "100%")
),
shiny::column(4, shiny::tableOutput(id))
)
)
}
levelsInput <- function(id, label, placeholder, blank_if_none = TRUE) {
shiny::fluidRow(
shiny::column(8,
shiny::textInput(id,
paste0("Indicate the ",
label,
", separated by commas",
ifelse(blank_if_none, " (leave blank if none).", ".")),
placeholder = placeholder,
width = "90%")),
shiny::column(4, shiny::tableOutput(id)))
}
dataset_panel <- function(title, id) shiny::tabPanel(title,
shiny::downloadButton(paste0("dl_", id)),
shiny::dataTableOutput(id))
mainButton <- function(id, label) shiny::actionButton(id, label, width = "100%", class = "btn btn-success")
warning_perso <- function(condition, text) {
shiny::conditionalPanel(condition = condition,
shiny::div(class = "panel panel-warning",
shiny::div(class = "panel-heading",
shiny::h3(class = "panel-title", text))))
}
warning_wd <- function() warning_perso("document.getElementById('button_clean').disabled",
"Please select the project directory.")
fun_waiter <- function(h1) waiter::Waiter$new(html = shiny::div(shiny::h1(h1, style = "color: white"),
waiter::spin_double_bounce()),
color = "#4CAF50")
ui <- shiny::tagList(
shinyjs::useShinyjs(),
waiter::use_waiter(),
shiny::navbarPage(
theme = shinythemes::shinytheme("paper"),
"mscleanr",
shiny::tabPanel(
"Project directory",
shiny::sidebarPanel(
shiny::h5("Project directory"),
"Select the directory containing the data you want to analyze.",
shiny::br(),
shiny::h5("Overwrite"),
"If checked, any existing file resulting from a previous analysis will be overwritten without asking."
),
shiny::mainPanel(
warning_wd(),
# project_directory,
shinyFiles::shinyDirButton("project_directory",
"Select the project directory",
"Select the project directory",
class = "btn btn-success",
style = "width: 100%"),
shiny::verbatimTextOutput("project_directory_path"),
# overwrite = TRUE)
shiny::checkboxInput("overwrite",
"Overwrite existing results?",
value = TRUE),
shiny::br(),
shiny::fluidRow(
shiny::fileInput("parameters",
paste0("Load saved parameters:"),
accept = ".csv",
placeholder = " No parameters file selected",
width = "100%"),
shiny::tableOutput("parameters_feedback")
)
)
),
shiny::tabPanel(
"Clean MS-DIAL data",
shiny::sidebarPanel(
shiny::h6("Filter peaks according to user parameters and optionnaly combine positive and negative feature lists if available."),
shiny::br(),
shiny::h5("Generic filters"),
"Check which filters you want to use to clean your MS data.",
shiny::h5("Clusterization deltas"),
"Indicates the acceptable retention time and mass differences to consider for feature relationships clusterization.",
shiny::h5("Clusterization options"),
"You can choose to use the Pearson correlation between peaks to extend cluster size",
shiny::h5("(Optional) Reference files"),
"Optionally, you can use your own files for adducts and neutral losses.
See the documentation for more information."
),
shiny::mainPanel(
warning_wd(),
shiny::fluidRow(
# filter_blk = TRUE,
# filter_mz = TRUE,
# filter_rsd = TRUE,
# filter_rmd = TRUE,
shiny::column(3,
shiny::checkboxGroupInput("filters",
"What filters to use?",
choiceNames = c("Blank ratio",
"Incorrect Mass",
"Relative Standard Deviation",
"Relative Mass Defect"),
choiceValues = c("filter_blk",
"filter_mz",
"filter_rsd",
"filter_rmd"),
selected = c("filter_blk",
"filter_mz",
"filter_rsd",
"filter_rmd"))
),
shiny::column(3,
shiny::conditionalPanel(
condition = "input.filters.includes('filter_blk')", # JS expression
# filter_blk_threshold = 0.8,
shiny::numericInput("filter_blk_threshold",
"Minimum blank ratio",
value = 0.8,
min = 0,
max = 1,
step = 0.05),
# filter_blk_ghost_peaks = TRUE,
shiny::checkboxInput("filter_blk_ghost_peaks",
"Delete ghost peaks?",
value = TRUE)
)
),
# filter_rsd_threshold = 30,
shiny::column(3,
shiny::conditionalPanel(
condition = "input.filters.includes('filter_rsd')", # JS expression
shiny::numericInput("filter_rsd_threshold",
"Maximum RSD",
value = 30,
min = 0,
max = 100,
step = 5)
)
),
# filter_rmd_range = c(50, 3000),
shiny::column(3,
shiny::conditionalPanel(
condition = "input.filters.includes('filter_rmd')", # JS expression
shiny::numericInput("filter_rmd_range_min",
"Minimum RMD",
value = 50,
min = 0,
max = 5000,
step = 5),
shiny::numericInput("filter_rmd_range_max",
"Maximum RMD",
value = 3000,
min = 5000,
max = 10000,
step = 5)
)
)
),
shiny::hr(),
shiny::fluidRow(
# threshold_mz = 0.005,
shiny::column(4,
shiny::numericInput("threshold_mz",
"Maximum mass difference",
value = 0.005,
min = 0,
step = 0.001)
),
# threshold_rt = 0.025,
shiny::column(4,
shiny::numericInput("threshold_rt",
"Maximum retention time difference",
value = 0.025,
min = 0,
step = 0.001)
)
),
shiny::hr(),
shiny::fluidRow(
# compute_pearson_correlation = TRUE,
shiny::column(4,
shiny::checkboxInput("compute_pearson_correlation",
"Use Pearson correlation to compute clusters?",
value = TRUE),
shiny::br()
),
# pearson_correlation_threshold = 0.8,
shiny::column(4,
shiny::conditionalPanel(
condition = "input.compute_pearson_correlation", # JS expression
shiny::numericInput("pearson_correlation_threshold",
"Minimum correlation",
value = 0.8,
min = 0,
max = 1,
step = 0.05)
)
),
# pearson_p_value = 0.05,
shiny::column(4,
shiny::conditionalPanel(
condition = "input.compute_pearson_correlation",
shiny::selectInput("pearson_p_value",
"Maximum p-value",
c("No filtering"=1, 0.05, 0.01),
multiple = FALSE,
selected = 0.05)
)
)
),
shiny::hr(),
"You can optionally import personal reference files for adducts and neutral losses.",
"By default, data displayed in the Datasets tab will be used.",
shiny::checkboxInput("user_refs",
"Use personal reference files?",
value = FALSE),
shiny::conditionalPanel(
condition = "input.user_refs",
shiny::tabsetPanel(
# user_pos_adducts_refs = NA,
referenceInput("user_pos_adducts_refs", "Positive adducts"),
# user_neg_adducts_refs = NA,
referenceInput("user_neg_adducts_refs", "Negative adducts"),
# user_pos_neutral_refs = NA,
referenceInput("user_pos_neutral_refs", "Neutral losses in positive mode"),
# user_neg_neutral_refs = NA,
referenceInput("user_neg_neutral_refs", "Neutral losses in negative mode")
)
),
mainButton("button_clean", "Clean MS-DIAL data"),
shiny::verbatimTextOutput("clean")
)
),
shiny::tabPanel(
"Keep top peaks by cluster",
shiny::sidebarPanel(
shiny::h6("Keeps only the top peaks by cluster."),
shiny::br(),
shiny::h5("Selection mode"),
"Peaks selected can be the most intense (intensity), most connected (degree), or both.",
"If there are ties, the number of peaks selected can be greater than the number requested by the user.",
shiny::h5("Exporting filtered peaks"),
"Copy MAT files corresponding to the selected peaks in a new folder for a faster analysis in MS-FINDER."
),
shiny::mainPanel(
warning_wd(),
warning_perso("!(input.selection_criteria.includes('intensity') | input.selection_criteria.includes('degree'))",
"Please select at least one selection criterion."),
shiny::fluidRow(
# selection_criterion = "degree",
shiny::column(4,
shiny::checkboxGroupInput("selection_criteria",
"Selection criteria",
choiceNames = c("Intensity (most intense peaks)",
"Degree (most connected peaks)"),
choiceValues = c("intensity", "degree"),
selected = "intensity")
),
# n = 1,
shiny::column(4,
shiny::numericInput("n",
"Number of peaks to keep (by cluster and by method)",
value = 1,
min = 1,
step = 1)
),
# export_filtered_peaks = TRUE,
shiny::column(4,
shiny::checkboxInput("export_filtered_peaks",
"Export final peaks in a new folder?",
value = TRUE)
)
),
mainButton("button_keep", "Keep top peaks by cluster"),
shiny::verbatimTextOutput("keep")
)
),
shiny::tabPanel(
"Launch MS-FINDER annotation",
shiny::sidebarPanel(
shiny::h6("Concatenate MS-FINDER annotations to the final filtered peak list."),
shiny::br(),
shiny::h5("Compound levels prioritization"),
"The list of compound levels to consider, in the given order (from more important to least important).A Compound_level column must be filled in your in-house database used for in silico based annotation",
shiny::h5("Biosource levels prioritization"),
"The list of biosource levels to consider, in the given order (from more important to least important).",
"They must correspond to the folders containing MS-FINDER output annotation files in your project directory.",
"The level 'generic' is always added as the last biosource level considered.",
shiny::h5("Levels scores"),
"A list of levels names and their corresponding multiplier to tune final annotation scores."
),
shiny::mainPanel(
warning_wd(),
shiny::checkboxInput("other_peaks_warning",
"Generate a warning if there are several possible annotated peaks in the cluster?",
value = FALSE,
width = "90%"),
shiny::checkboxInput("score_only",
"Select the best annotation for each peak based only on MS-FINDER scores?",
value = FALSE,
width = "90%"),
shiny::br(),
# biosoc_levels = c("genre", "family"),
levelsInput("biosoc_levels",
"biosource levels in your annotation process",
"genus,family,...",
blank_if_none = FALSE),
shiny::conditionalPanel(
condition = "!input.score_only",
# compound_levels = c("1a", "1b"),
levelsInput("compound_levels",
"compound levels in your annotation files",
"direct,extension,..."),
# levels_scores = list("1a" = 2, "1b" = 1.5, "genre" = 2, "family" = 1.5),
levelsInput("levels_scores",
"scores multipliers associated to your compound or biosource levels",
"direct:2,genus:2,family:1.5,...")
),
mainButton("button_launch", "Launch MS-FINDER annotation"),
shiny::verbatimTextOutput("launch")
)
),
shiny::tabPanel(
"Convert annotated peaks to MSP files",
shiny::sidebarPanel(
shiny::h6("Convert the final CSV file post annotations to MSP format."),
shiny::br(),
shiny::h5("Minimum score"),
"Minimum annotation score needed to export peaks to the MSP files."
),
shiny::mainPanel(
warning_wd(),
shiny::checkboxInput("all",
"Export all peaks to MSP files?",
value = TRUE),
shiny::conditionalPanel(
condition = "!input.all",
shiny::numericInput("min_score",
"Minimum score to convert peak to MSP",
value = 10,
min = 0)
),
mainButton("button_convert", "Convert peaks to MSP files"),
shiny::verbatimTextOutput("convert")
)
),
# shiny::tabPanel(
# "PDA / CAD",
#
# shiny::sidebarPanel(
# shiny::h6("Clean CAD/PDA peaks based on manually extracted data and joins them with the annotated MS data."),
# shiny::br(),
# shiny::h5("Start row"),
# "The row number of the start of the CAD/PDA RT and intensities in your manually extracted file.",
# shiny::h5("RT delta"),
# "Positive or negative value indicating the difference in RT between the CAD/PDA peaks and MS peaks.",
# shiny::h5("RT threshold"),
# "The difference in RT between the CAD/PDA peaks and MS peaks tolerated to join them.",
# shiny::h5("Pearson correlation and p-value thresholds."),
# "Links having a too low or a non-significative Pearson correlation will be ignored."
# ),
#
# shiny::mainPanel(
# warning_wd(),
# shiny::selectInput("mode",
# shiny::h5("Mode"),
# c("CAD", "PDA"),
# multiple = FALSE,
# selected = "CAD",
# width = "100%"),
# shiny::br(),
# shiny::column(4,
# # clean_cad_pda_peaks("PDA", data_start_row = 5, plot_spectra = FALSE, plot_ROI = FALSE)
# shiny::h5("Clean peaks"),
# shiny::numericInput("data_start_row",
# "Start row",
# value = 5,
# min = 1,
# step = 1),
# shiny::checkboxInput("plot_spectra",
# "Export spectra plot?",
# value = FALSE),
# shiny::checkboxInput("plot_ROI",
# "Export ROI plot?",
# value = FALSE),
# mainButton("button_cad_clean", "Launch cleaning"),
# shiny::verbatimTextOutput("cad_clean")
# ),
# shiny::column(8,
# shiny::h5("Merge with MS peaks"),
# shiny::fluidRow(
# shiny::column(6,
# shiny::numericInput("delta_rt",
# "RT delta",
# value = 0,
# step = 0.01),
# shiny::helpText("RT(CAD/PDA) + RT_delta = RT(MS)")
# ),
# shiny::column(6,
# shiny::numericInput("cad_threshold_rt",
# "RT threshold",
# value = 0.05,
# min = 0,
# step = 0.01)
# )
# ),
# shiny::fluidRow(
# shiny::column(6,
# shiny::numericInput("cad_pearson_correlation_threshold",
# "Minimum Pearson correlation",
# value = 0.8,
# min = 0,
# max = 1,
# step = 0.05)
# ),
# shiny::column(6,
# shiny::selectInput("cad_pearson_p_value",
# "Maximum Pearson p-value",
# c("No filtering"=1, 0.05, 0.01),
# multiple = FALSE,
# selected = 0.05)
# )
# ),
# mainButton("button_cad_merge", "Launch merging"),
# shiny::verbatimTextOutput("cad_merge")
# )
# )
# ),
shiny::tabPanel(
"Datasets",
shiny::sidebarPanel(
shiny::h6("Datasets used by default during the analysis if none are provided by the user."),
shiny::br(),
shiny::h5("Adducts"),
"Mass differences of adducts with the original compound M, in positive and negative mode.",
shiny::h5("Neutral losses"),
"Masses of neutral losses, in positive and negative mode.",
shiny::h5("Isotopes"),
"Masses of isotopes."
),
shiny::mainPanel(
shiny::tabsetPanel(
dataset_panel("Positive adducts", "adducts_pos"),
dataset_panel("Negative adducts", "adducts_neg"),
dataset_panel("Neutral losses in positive mode", "nl_pos"),
dataset_panel("Neutral losses in negative mode", "nl_neg"),
shiny::tabPanel("Isotopes", shiny::dataTableOutput("isotopes"))
)
)
)
)
)
server <- function(input, output, session) {
# Project directory
volumes <- c(Home = fs::path_home(), shinyFiles::getVolumes()())
shinyFiles::shinyDirChoose(input,
"project_directory",
roots = volumes,
session = session)
output$project_directory_path <- shiny::renderPrint({
if (is.integer(input$project_directory)) {
cat("No project directory has been selected")
shinyjs::disable("button_clean")
shinyjs::disable("button_keep")
shinyjs::disable("button_launch")
shinyjs::disable("button_convert")
shinyjs::disable("button_cad_clean")
shinyjs::disable("button_cad_merge")
} else {
path <- shinyFiles::parseDirPath(volumes, input$project_directory)
assign("analysis_directory", path, envir = mscleanrCache)
cat(path)
shinyjs::enable("button_clean")
shinyjs::enable("button_keep")
shinyjs::enable("button_launch")
shinyjs::enable("button_convert")
shinyjs::enable("button_cad_clean")
shinyjs::enable("button_cad_merge")
}
})
shiny::observeEvent(input$overwrite, { assign("overwrite", input$overwrite, envir = mscleanrCache) })
# Parameters
shinyjs::runjs("$('#parameters').parent().addClass('btn-sm');")
shinyjs::runjs("$('#parameters').parent().parent().parent().parent().parent().addClass('well').addClass('well-sm');")
output$parameters_feedback <- shiny::renderTable({
shiny::req(input$parameters)
params <- vroom::vroom(input$parameters$datapath)
shiny::validate(shiny::need(ncol(params) == 3 & "PARAMETER" %in% names(params) & "VALUE" %in% names(params),
"Incorrect parameters file."))
params$loaded <- FALSE
# Ignored: *_refs, Package version
params <- params[!endsWith(params$PARAMETER, "_refs") & params$PARAMETER != "Package version",]
# checkboxInput: filter_blk_ghost_peaks, compute_pearson_correlation
booleans <- params[params$VALUE %in% c("TRUE", "FALSE"),]$PARAMETER
for (pb in booleans[booleans %in% names(input)]) {
try({
shiny::updateCheckboxInput(session, pb, value = as.logical(params[params$PARAMETER == pb,]$VALUE))
params[params$PARAMETER == pb,]$loaded <- TRUE
})
}
# filter_blk, filter_mz, filter_rsd, filter_rmd -> strings in checkboxGroupInput "filters"
try({
group_filters <- booleans[!(booleans %in% names(input))
& booleans %in% c("filter_blk", "filter_mz", "filter_rsd", "filter_rmd")]
shiny::updateCheckboxGroupInput(session, "filters", selected = group_filters)
for (pf in group_filters) params[params$PARAMETER == pf,]$loaded <- TRUE
})
# selectInput : pearson_p_value
if ("pearson_p_value" %in% params$PARAMETER) {
try({
choices <- c(1, 0.05, 0.01)
diff <- as.numeric(params[params$PARAMETER == "pearson_p_value",]$VALUE) - unlist(choices)
closest <- choices[which.min(abs(diff))]
shiny::updateSelectInput(session, "pearson_p_value", selected = closest)
params[params$PARAMETER == "pearson_p_value",]$loaded <- TRUE
})
}
# filter_rmd_range -> 2 numericInput "filter_rmd_range_min" and "filter_rmd_range_max"
if ("filter_rmd_range" %in% params$PARAMETER) {
try({
str_range <- as.character(params[params$PARAMETER == "filter_rmd_range",]$VALUE)
str_range <- as.numeric(unlist(strsplit(substr(str_range, 2, nchar(str_range)-1), ",")))
shiny::updateNumericInput(session, "filter_rmd_range_min", value = str_range[1])
shiny::updateNumericInput(session, "filter_rmd_range_max", value = str_range[2])
params[params$PARAMETER == "filter_rmd_range",]$loaded <- TRUE
})
}
# selection_criterion -> strings in checkboxGroupInput "selection_criteria" ("intensity", "degree")
if ("selection_criterion" %in% params$PARAMETER) {
if (params[params$PARAMETER == "selection_criterion",]$VALUE %in% c("both", "intensity", "degree")) {
try({
if (params[params$PARAMETER == "selection_criterion",]$VALUE == "both") {
shiny::updateCheckboxGroupInput(session, "selection_criteria", selected = c("intensity", "degree"))
} else {
shiny::updateCheckboxGroupInput(session, "selection_criteria",
selected = c(as.character(params[params$PARAMETER == "selection_criterion",]$VALUE)))
}
params[params$PARAMETER == "selection_criterion",]$loaded <- TRUE
})
}
}
# textInput: compound_levels, biosoc_levels, levels_scores
for (pl in c("compound_levels", "biosoc_levels", "levels_scores")) {
try({
if (pl %in% params$PARAMETER & params[params$PARAMETER == pl,]$VALUE != "[]") {
str_list <- as.character(params[params$PARAMETER == pl,]$VALUE)
str_list <- substr(str_list, 2, nchar(str_list)-1)
shiny::updateTextInput(session, pl, value = str_list)
}
params[params$PARAMETER == pl,]$loaded <- TRUE
})
}
# numericInput
for (pn in c("filter_blk_threshold", "filter_rsd_threshold", "threshold_mz", "threshold_rt",
"pearson_correlation_threshold", "n", "min_score")) {
try({
shiny::updateNumericInput(session, pn, value = as.numeric(params[params$PARAMETER == pn,]$VALUE))
params[params$PARAMETER == pn,]$loaded <- TRUE
})
}
# Feedback for user
params
})
# References
user_ref_generic <- function(ref_file) {
shiny::req(ref_file)
data <- vroom::vroom(ref_file$datapath)
shiny::validate(shiny::need(ncol(data) == 2, "Please upload a reference file with 2 columns (see documentation)."))
return(data)
}
user_pos_adducts_refs <- shiny::reactive({ shiny::req(user_ref_generic(input$user_pos_adducts_refs))
user_ref_generic(input$user_pos_adducts_refs) })
user_neg_adducts_refs <- shiny::reactive({ shiny::req(user_ref_generic(input$user_neg_adducts_refs))
user_ref_generic(input$user_neg_adducts_refs) })
user_pos_neutral_refs <- shiny::reactive({ shiny::req(user_ref_generic(input$user_pos_neutral_refs))
user_ref_generic(input$user_pos_neutral_refs) })
user_neg_neutral_refs <- shiny::reactive({ shiny::req(user_ref_generic(input$user_neg_neutral_refs))
user_ref_generic(input$user_neg_neutral_refs) })
preview_ref <- function(ref_data) rbind(utils::head(ref_data, 2),
stats::setNames(data.frame(a = "...", b = "..."), names(ref_data)))
output$user_pos_adducts_refs <- shiny::renderTable(preview_ref(user_pos_adducts_refs()))
output$user_neg_adducts_refs <- shiny::renderTable(preview_ref(user_neg_adducts_refs()))
output$user_pos_neutral_refs <- shiny::renderTable(preview_ref(user_pos_neutral_refs()))
output$user_neg_neutral_refs <- shiny::renderTable(preview_ref(user_neg_neutral_refs()))
# clean_msdial_data
clean_params <- shiny::eventReactive(input$button_clean, {
check_ref <- function(ref) {
tryCatch({ shiny::validate(shiny::need(ref, "not found")) ; return(ref) },
error = function(e) {})
}
get_ref <- function(ref) if(is.null(check_ref(ref))) NA else as.data.frame(ref)
list(
filter_blk = "filter_blk" %in% input$filters,
filter_blk_threshold = input$filter_blk_threshold,
filter_blk_ghost_peaks = input$filter_blk_ghost_peaks,
filter_mz = "filter_mz" %in% input$filters,
filter_rsd = "filter_rsd" %in% input$filters,
filter_rsd_threshold = input$filter_rsd_threshold,
filter_rmd = "filter_rmd" %in% input$filters,
filter_rmd_range = c(input$filter_rmd_range_min, input$filter_rmd_range_max),
threshold_mz = input$threshold_mz,
threshold_rt = input$threshold_rt,
compute_pearson_correlation = input$compute_pearson_correlation,
pearson_correlation_threshold = input$pearson_correlation_threshold,
pearson_p_value = input$pearson_p_value,
user_pos_adducts_refs = get_ref(user_pos_adducts_refs()),
user_neg_adducts_refs = get_ref(user_neg_adducts_refs()),
user_pos_neutral_refs = get_ref(user_pos_neutral_refs()),
user_neg_neutral_refs = get_ref(user_neg_neutral_refs())
)
})
output$clean <- shiny::renderPrint({
shiny::req(input$button_clean > 0)
shiny::req(exists("analysis_directory", envir = mscleanrCache))
shiny::req(clean_params())
waiter <- fun_waiter("Cleaning data...")
waiter$show()
tryCatch(
do.call(clean_msdial_data, clean_params()),
finally = waiter$hide()
)
})
# keep_top_peaks
keep_params <- shiny::eventReactive(input$button_keep, {
shiny::req("degree" %in% input$selection_criteria | "intensity" %in% input$selection_criteria)
if ("degree" %in% input$selection_criteria) {
if ("intensity" %in% input$selection_criteria) mode <- "both"
else mode <- "degree"
} else mode <- "intensity"
list(selection_criterion = mode,
n = input$n,
export_filtered_peaks = input$export_filtered_peaks)
})
output$keep <- shiny::renderPrint({
shiny::req(input$button_keep > 0)
shiny::req(exists("analysis_directory", envir = mscleanrCache))
shiny::req(keep_params())
waiter <- fun_waiter("Keeping only selected peaks...")
waiter$show()
tryCatch(
do.call(keep_top_peaks, keep_params()),
finally = waiter$hide()
)
})
# launch_msfinder_annotation
clean <- function(l) {
tmp <- unlist(lapply(strsplit(l, ",", fixed = TRUE), function(x) trimws(x)))
tmp <- tmp[tmp != ""]
if (length(tmp) == 0) return(NULL) else return(unique(tmp))
}
compounds <- shiny::reactive(clean(input$compound_levels))
biosoc <- shiny::reactive({
tmp <- clean(input$biosoc_levels)
if (is.null(tmp)) c("generic") else tmp
})
scores <- shiny::reactive({
tmp_scores <- lapply(clean(input$levels_scores), strsplit, split = ":", fixed = TRUE)
if (length(tmp_scores) == 0) scores <- NULL
else {
scores <- list()
for (x in tmp_scores) scores[x[[1]][1]] <- ifelse(is.na(x[[1]][2]) | x[[1]][2] == "",
1,
as.numeric(x[[1]][2]))
}
scores
})
output$compound_levels <- shiny::renderTable({
shiny::req(compounds())
data.frame("Rank" = c(1:length(compounds())), "Compound level" = compounds())
})
output$biosoc_levels <- shiny::renderTable(
data.frame("Rank" = c(1:length(unique(c(biosoc(), "generic")))),
"Biosource level" = unique(c(biosoc(), "generic")))
)
output$levels_scores <- shiny::renderTable(
data.frame(Level = names(scores()), Multiplier = unlist(scores(), use.names = FALSE))
)
launch_params <- shiny::eventReactive(input$button_launch, {
list(compound_levels = compounds(),
biosoc_levels = biosoc(),
levels_scores = scores(),
score_only = input$score_only,
other_peaks_warning = input$other_peaks_warning)
})
output$launch <- shiny::renderPrint({
shiny::req(input$button_launch > 0)
shiny::req(exists("analysis_directory", envir = mscleanrCache))
shiny::req(launch_params())
waiter <- fun_waiter("Annotating peaks with MS-FINDER data...")
waiter$show()
tryCatch(
do.call(launch_msfinder_annotation, launch_params()),
finally = waiter$hide()
)
})
# convert_csv_to_msp
convert_params <- shiny::eventReactive(input$button_convert, list(all = input$all,
min_score = input$min_score))
output$convert <- shiny::renderPrint({
shiny::req(input$button_convert > 0)
shiny::req(exists("analysis_directory", envir = mscleanrCache))
shiny::req(convert_params())
waiter <- fun_waiter("Converting peaks to MSP...")
waiter$show()
tryCatch(
do.call(convert_csv_to_msp, convert_params()),
finally = waiter$hide()
)
})
# CAD/PDA
# cad_clean_params <- shiny::eventReactive(input$button_cad_clean, {
# list(mode = input$mode,
# data_start_row = input$data_start_row,
# plot_spectra = input$plot_spectra,
# plot_ROI = input$plot_ROI)
# })
#
# output$cad_clean <- shiny::renderPrint({
# shiny::req(input$button_cad_clean > 0)
# shiny::req(exists("analysis_directory", envir = mscleanrCache))
# shiny::req(cad_clean_params())
# waiter <- fun_waiter(paste("Cleaning", cad_clean_params()$mode, "peaks..."))
# waiter$show()
# tryCatch(
# do.call(clean_cad_pda_peaks, cad_clean_params()),
# finally = waiter$hide()
# )
# })
#
# cad_merge_params <- shiny::eventReactive(input$button_cad_merge, {
# list(mode = input$mode,
# delta_rt = input$delta_rt,
# threshold_rt = input$cad_threshold_rt,
# pearson_correlation_threshold = input$cad_pearson_correlation_threshold,
# pearson_p_value = input$cad_pearson_p_value)
# })
#
# output$cad_merge <- shiny::renderPrint({
# shiny::req(input$button_cad_merge > 0)
# shiny::req(exists("analysis_directory", envir = mscleanrCache))
# shiny::req(cad_merge_params())
# waiter <- fun_waiter(paste("Merging", cad_merge_params()$mode, "and MS peaks..."))
# waiter$show()
# tryCatch(
# do.call(launch_cad_fusion, cad_merge_params()),
# finally = waiter$hide()
# )
# })
# Datasets
output$adducts_pos <- shiny::renderDataTable(mass_adducts_pos, escape = FALSE, options = list(pageLength = 10))
output$adducts_neg <- shiny::renderDataTable(mass_adducts_neg, escape = FALSE, options = list(pageLength = 10))
output$nl_pos <- shiny::renderDataTable(mass_neutral_loss_pos, escape = FALSE, options = list(pageLength = 10))
output$nl_neg <- shiny::renderDataTable(mass_neutral_loss_neg, escape = FALSE, options = list(pageLength = 10))
output$isotopes <- shiny::renderDataTable(mass_isotopes, escape = FALSE, options = list(pageLength = 10))
dl <- function(data, filename) shiny::downloadHandler(filename = filename,
content = function(file) vroom::vroom_write(data, file, delim = ","))
output$dl_adducts_pos <- dl(mass_adducts_pos, "positive_adducts.csv")
output$dl_adducts_neg <- dl(mass_adducts_neg, "negative_adducts.csv")
output$dl_nl_pos <- dl(mass_neutral_loss_pos, "positive_neutral_losses.csv")
output$dl_nl_neg <- dl(mass_neutral_loss_neg, "negative_neutral_losses.csv")
shiny::onStop(function() { assign("shiny_running", FALSE, envir = mscleanrCache) })
}
shiny::shinyApp(ui = ui,
server = server,
onStart = function() { assign("shiny_running", TRUE, envir = mscleanrCache) })
}
eMetaboHUB/MS-CleanR documentation built on Jan. 3, 2024, 8:55 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Defines functions runGUI
Documented in runGUI
#' Launch the Graphical User Interface (GUI)
#'
#' @export
runGUI <- function() {
# datasets loading
utils::data("mass_neutral_loss_neg", package="mscleanr", envir=environment())
mass_neutral_loss_neg <- get("mass_neutral_loss_neg", envir=environment()) # redundant, but makes the syntax checker happy
utils::data("mass_neutral_loss_pos", package="mscleanr", envir=environment())
mass_neutral_loss_pos <- get("mass_neutral_loss_pos", envir=environment()) # redundant, but makes the syntax checker happy
utils::data("mass_adducts_neg", package="mscleanr", envir=environment())
mass_adducts_neg <- get("mass_adducts_neg", envir=environment()) # redundant, but makes the syntax checker happy
utils::data("mass_adducts_pos", package="mscleanr", envir=environment())
mass_adducts_pos <- get("mass_adducts_pos", envir=environment()) # redundant, but makes the syntax checker happy
utils::data("mass_isotopes", package="mscleanr", envir=environment())
mass_isotopes <- get("mass_isotopes", envir=environment()) # redundant, but makes the syntax checker happy
# personalized elements
referenceInput <- function(id, text) {
shiny::tabPanel(text,
shiny::fluidRow(
shiny::column(8,
shiny::fileInput(id,
paste0("(OPTIONAL) Upload a CSV reference file for ", tolower(text), "."),
accept = ".csv",
placeholder = " No file selected",
width = "100%")
),
shiny::column(4, shiny::tableOutput(id))
)
)
}
levelsInput <- function(id, label, placeholder, blank_if_none = TRUE) {
shiny::fluidRow(
shiny::column(8,
shiny::textInput(id,
paste0("Indicate the ",
label,
", separated by commas",
ifelse(blank_if_none, " (leave blank if none).", ".")),
placeholder = placeholder,
width = "90%")),
shiny::column(4, shiny::tableOutput(id)))
}
dataset_panel <- function(title, id) shiny::tabPanel(title,
shiny::downloadButton(paste0("dl_", id)),
shiny::dataTableOutput(id))
mainButton <- function(id, label) shiny::actionButton(id, label, width = "100%", class = "btn btn-success")
warning_perso <- function(condition, text) {
shiny::conditionalPanel(condition = condition,
shiny::div(class = "panel panel-warning",
shiny::div(class = "panel-heading",
shiny::h3(class = "panel-title", text))))
}
warning_wd <- function() warning_perso("document.getElementById('button_clean').disabled",
"Please select the project directory.")
fun_waiter <- function(h1) waiter::Waiter$new(html = shiny::div(shiny::h1(h1, style = "color: white"),
waiter::spin_double_bounce()),
color = "#4CAF50")
ui <- shiny::tagList(
shinyjs::useShinyjs(),
waiter::use_waiter(),
shiny::navbarPage(
theme = shinythemes::shinytheme("paper"),
"mscleanr",
shiny::tabPanel(
"Project directory",
shiny::sidebarPanel(
shiny::h5("Project directory"),
"Select the directory containing the data you want to analyze.",
shiny::br(),
shiny::h5("Overwrite"),
"If checked, any existing file resulting from a previous analysis will be overwritten without asking."
),
shiny::mainPanel(
warning_wd(),
# project_directory,
shinyFiles::shinyDirButton("project_directory",
"Select the project directory",
"Select the project directory",
class = "btn btn-success",
style = "width: 100%"),
shiny::verbatimTextOutput("project_directory_path"),
# overwrite = TRUE)
shiny::checkboxInput("overwrite",
"Overwrite existing results?",
value = TRUE),
shiny::br(),
shiny::fluidRow(
shiny::fileInput("parameters",
paste0("Load saved parameters:"),
accept = ".csv",
placeholder = " No parameters file selected",
width = "100%"),
shiny::tableOutput("parameters_feedback")
)
)
),
shiny::tabPanel(
"Clean MS-DIAL data",
shiny::sidebarPanel(
shiny::h6("Filter peaks according to user parameters and optionnaly combine positive and negative feature lists if available."),
shiny::br(),
shiny::h5("Generic filters"),
"Check which filters you want to use to clean your MS data.",
shiny::h5("Clusterization deltas"),
"Indicates the acceptable retention time and mass differences to consider for feature relationships clusterization.",
shiny::h5("Clusterization options"),
"You can choose to use the Pearson correlation between peaks to extend cluster size",
shiny::h5("(Optional) Reference files"),
"Optionally, you can use your own files for adducts and neutral losses.
See the documentation for more information."
),
shiny::mainPanel(
warning_wd(),
shiny::fluidRow(
# filter_blk = TRUE,
# filter_mz = TRUE,
# filter_rsd = TRUE,
# filter_rmd = TRUE,
shiny::column(3,
shiny::checkboxGroupInput("filters",
"What filters to use?",
choiceNames = c("Blank ratio",
"Incorrect Mass",
"Relative Standard Deviation",
"Relative Mass Defect"),
choiceValues = c("filter_blk",
"filter_mz",
"filter_rsd",
"filter_rmd"),
selected = c("filter_blk",
"filter_mz",
"filter_rsd",
"filter_rmd"))
),
shiny::column(3,
shiny::conditionalPanel(
condition = "input.filters.includes('filter_blk')", # JS expression
# filter_blk_threshold = 0.8,
shiny::numericInput("filter_blk_threshold",
"Minimum blank ratio",
value = 0.8,
min = 0,
max = 1,
step = 0.05),
# filter_blk_ghost_peaks = TRUE,
shiny::checkboxInput("filter_blk_ghost_peaks",
"Delete ghost peaks?",
value = TRUE)
)
),
# filter_rsd_threshold = 30,
shiny::column(3,
shiny::conditionalPanel(
condition = "input.filters.includes('filter_rsd')", # JS expression
shiny::numericInput("filter_rsd_threshold",
"Maximum RSD",
value = 30,
min = 0,
max = 100,
step = 5)
)
),
# filter_rmd_range = c(50, 3000),
shiny::column(3,
shiny::conditionalPanel(
condition = "input.filters.includes('filter_rmd')", # JS expression
shiny::numericInput("filter_rmd_range_min",
"Minimum RMD",
value = 50,
min = 0,
max = 5000,
step = 5),
shiny::numericInput("filter_rmd_range_max",
"Maximum RMD",
value = 3000,
min = 5000,
max = 10000,
step = 5)
)
)
),
shiny::hr(),
shiny::fluidRow(
# threshold_mz = 0.005,
shiny::column(4,
shiny::numericInput("threshold_mz",
"Maximum mass difference",
value = 0.005,
min = 0,
step = 0.001)
),
# threshold_rt = 0.025,
shiny::column(4,
shiny::numericInput("threshold_rt",
"Maximum retention time difference",
value = 0.025,
min = 0,
step = 0.001)
)
),
shiny::hr(),
shiny::fluidRow(
# compute_pearson_correlation = TRUE,
shiny::column(4,
shiny::checkboxInput("compute_pearson_correlation",
"Use Pearson correlation to compute clusters?",
value = TRUE),
shiny::br()
),
# pearson_correlation_threshold = 0.8,
shiny::column(4,
shiny::conditionalPanel(
condition = "input.compute_pearson_correlation", # JS expression
shiny::numericInput("pearson_correlation_threshold",
"Minimum correlation",
value = 0.8,
min = 0,
max = 1,
step = 0.05)
)
),
# pearson_p_value = 0.05,
shiny::column(4,
shiny::conditionalPanel(
condition = "input.compute_pearson_correlation",
shiny::selectInput("pearson_p_value",
"Maximum p-value",
c("No filtering"=1, 0.05, 0.01),
multiple = FALSE,
selected = 0.05)
)
)
),
shiny::hr(),
"You can optionally import personal reference files for adducts and neutral losses.",
"By default, data displayed in the Datasets tab will be used.",
shiny::checkboxInput("user_refs",
"Use personal reference files?",
value = FALSE),
shiny::conditionalPanel(
condition = "input.user_refs",
shiny::tabsetPanel(
# user_pos_adducts_refs = NA,
referenceInput("user_pos_adducts_refs", "Positive adducts"),
# user_neg_adducts_refs = NA,
referenceInput("user_neg_adducts_refs", "Negative adducts"),
# user_pos_neutral_refs = NA,
referenceInput("user_pos_neutral_refs", "Neutral losses in positive mode"),
# user_neg_neutral_refs = NA,
referenceInput("user_neg_neutral_refs", "Neutral losses in negative mode")
)
),
mainButton("button_clean", "Clean MS-DIAL data"),
shiny::verbatimTextOutput("clean")
)
),
shiny::tabPanel(
"Keep top peaks by cluster",
shiny::sidebarPanel(
shiny::h6("Keeps only the top peaks by cluster."),
shiny::br(),
shiny::h5("Selection mode"),
"Peaks selected can be the most intense (intensity), most connected (degree), or both.",
"If there are ties, the number of peaks selected can be greater than the number requested by the user.",
shiny::h5("Exporting filtered peaks"),
"Copy MAT files corresponding to the selected peaks in a new folder for a faster analysis in MS-FINDER."
),
shiny::mainPanel(
warning_wd(),
warning_perso("!(input.selection_criteria.includes('intensity') | input.selection_criteria.includes('degree'))",
"Please select at least one selection criterion."),
shiny::fluidRow(
# selection_criterion = "degree",
shiny::column(4,
shiny::checkboxGroupInput("selection_criteria",
"Selection criteria",
choiceNames = c("Intensity (most intense peaks)",
"Degree (most connected peaks)"),
choiceValues = c("intensity", "degree"),
selected = "intensity")
),
# n = 1,
shiny::column(4,
shiny::numericInput("n",
"Number of peaks to keep (by cluster and by method)",
value = 1,
min = 1,
step = 1)
),
# export_filtered_peaks = TRUE,
shiny::column(4,
shiny::checkboxInput("export_filtered_peaks",
"Export final peaks in a new folder?",
value = TRUE)
)
),
mainButton("button_keep", "Keep top peaks by cluster"),
shiny::verbatimTextOutput("keep")
)
),
shiny::tabPanel(
"Launch MS-FINDER annotation",
shiny::sidebarPanel(
shiny::h6("Concatenate MS-FINDER annotations to the final filtered peak list."),
shiny::br(),
shiny::h5("Compound levels prioritization"),
"The list of compound levels to consider, in the given order (from more important to least important).A Compound_level column must be filled in your in-house database used for in silico based annotation",
shiny::h5("Biosource levels prioritization"),
"The list of biosource levels to consider, in the given order (from more important to least important).",
"They must correspond to the folders containing MS-FINDER output annotation files in your project directory.",
"The level 'generic' is always added as the last biosource level considered.",
shiny::h5("Levels scores"),
"A list of levels names and their corresponding multiplier to tune final annotation scores."
),
shiny::mainPanel(
warning_wd(),
shiny::checkboxInput("other_peaks_warning",
"Generate a warning if there are several possible annotated peaks in the cluster?",
value = FALSE,
width = "90%"),
shiny::checkboxInput("score_only",
"Select the best annotation for each peak based only on MS-FINDER scores?",
value = FALSE,
width = "90%"),
shiny::br(),
# biosoc_levels = c("genre", "family"),
levelsInput("biosoc_levels",
"biosource levels in your annotation process",
"genus,family,...",
blank_if_none = FALSE),
shiny::conditionalPanel(
condition = "!input.score_only",
# compound_levels = c("1a", "1b"),
levelsInput("compound_levels",
"compound levels in your annotation files",
"direct,extension,..."),
# levels_scores = list("1a" = 2, "1b" = 1.5, "genre" = 2, "family" = 1.5),
levelsInput("levels_scores",
"scores multipliers associated to your compound or biosource levels",
"direct:2,genus:2,family:1.5,...")
),
mainButton("button_launch", "Launch MS-FINDER annotation"),
shiny::verbatimTextOutput("launch")
)
),
shiny::tabPanel(
"Convert annotated peaks to MSP files",
shiny::sidebarPanel(
shiny::h6("Convert the final CSV file post annotations to MSP format."),
shiny::br(),
shiny::h5("Minimum score"),
"Minimum annotation score needed to export peaks to the MSP files."
),
shiny::mainPanel(
warning_wd(),
shiny::checkboxInput("all",
"Export all peaks to MSP files?",
value = TRUE),
shiny::conditionalPanel(
condition = "!input.all",
shiny::numericInput("min_score",
"Minimum score to convert peak to MSP",
value = 10,
min = 0)
),
mainButton("button_convert", "Convert peaks to MSP files"),
shiny::verbatimTextOutput("convert")
)
),
# shiny::tabPanel(
# "PDA / CAD",
#
# shiny::sidebarPanel(
# shiny::h6("Clean CAD/PDA peaks based on manually extracted data and joins them with the annotated MS data."),
# shiny::br(),
# shiny::h5("Start row"),
# "The row number of the start of the CAD/PDA RT and intensities in your manually extracted file.",
# shiny::h5("RT delta"),
# "Positive or negative value indicating the difference in RT between the CAD/PDA peaks and MS peaks.",
# shiny::h5("RT threshold"),
# "The difference in RT between the CAD/PDA peaks and MS peaks tolerated to join them.",
# shiny::h5("Pearson correlation and p-value thresholds."),
# "Links having a too low or a non-significative Pearson correlation will be ignored."
# ),
#
# shiny::mainPanel(
# warning_wd(),
# shiny::selectInput("mode",
# shiny::h5("Mode"),
# c("CAD", "PDA"),
# multiple = FALSE,
# selected = "CAD",
# width = "100%"),
# shiny::br(),
# shiny::column(4,
# # clean_cad_pda_peaks("PDA", data_start_row = 5, plot_spectra = FALSE, plot_ROI = FALSE)
# shiny::h5("Clean peaks"),
# shiny::numericInput("data_start_row",
# "Start row",
# value = 5,
# min = 1,
# step = 1),
# shiny::checkboxInput("plot_spectra",
# "Export spectra plot?",
# value = FALSE),
# shiny::checkboxInput("plot_ROI",
# "Export ROI plot?",
# value = FALSE),
# mainButton("button_cad_clean", "Launch cleaning"),
# shiny::verbatimTextOutput("cad_clean")
# ),
# shiny::column(8,
# shiny::h5("Merge with MS peaks"),
# shiny::fluidRow(
# shiny::column(6,
# shiny::numericInput("delta_rt",
# "RT delta",
# value = 0,
# step = 0.01),
# shiny::helpText("RT(CAD/PDA) + RT_delta = RT(MS)")
# ),
# shiny::column(6,
# shiny::numericInput("cad_threshold_rt",
# "RT threshold",
# value = 0.05,
# min = 0,
# step = 0.01)
# )
# ),
# shiny::fluidRow(
# shiny::column(6,
# shiny::numericInput("cad_pearson_correlation_threshold",
# "Minimum Pearson correlation",
# value = 0.8,
# min = 0,
# max = 1,
# step = 0.05)
# ),
# shiny::column(6,
# shiny::selectInput("cad_pearson_p_value",
# "Maximum Pearson p-value",
# c("No filtering"=1, 0.05, 0.01),
# multiple = FALSE,
# selected = 0.05)
# )
# ),
# mainButton("button_cad_merge", "Launch merging"),
# shiny::verbatimTextOutput("cad_merge")
# )
# )
# ),
shiny::tabPanel(
"Datasets",
shiny::sidebarPanel(
shiny::h6("Datasets used by default during the analysis if none are provided by the user."),
shiny::br(),
shiny::h5("Adducts"),
"Mass differences of adducts with the original compound M, in positive and negative mode.",
shiny::h5("Neutral losses"),
"Masses of neutral losses, in positive and negative mode.",
shiny::h5("Isotopes"),
"Masses of isotopes."
),
shiny::mainPanel(
shiny::tabsetPanel(
dataset_panel("Positive adducts", "adducts_pos"),
dataset_panel("Negative adducts", "adducts_neg"),
dataset_panel("Neutral losses in positive mode", "nl_pos"),
dataset_panel("Neutral losses in negative mode", "nl_neg"),
shiny::tabPanel("Isotopes", shiny::dataTableOutput("isotopes"))
)
)
)
)
)
server <- function(input, output, session) {
# Project directory
volumes <- c(Home = fs::path_home(), shinyFiles::getVolumes()())
shinyFiles::shinyDirChoose(input,
"project_directory",
roots = volumes,
session = session)
output$project_directory_path <- shiny::renderPrint({
if (is.integer(input$project_directory)) {
cat("No project directory has been selected")
shinyjs::disable("button_clean")
shinyjs::disable("button_keep")
shinyjs::disable("button_launch")
shinyjs::disable("button_convert")
shinyjs::disable("button_cad_clean")
shinyjs::disable("button_cad_merge")
} else {
path <- shinyFiles::parseDirPath(volumes, input$project_directory)
assign("analysis_directory", path, envir = mscleanrCache)
cat(path)
shinyjs::enable("button_clean")
shinyjs::enable("button_keep")
shinyjs::enable("button_launch")
shinyjs::enable("button_convert")
shinyjs::enable("button_cad_clean")
shinyjs::enable("button_cad_merge")
}
})
shiny::observeEvent(input$overwrite, { assign("overwrite", input$overwrite, envir = mscleanrCache) })
# Parameters
shinyjs::runjs("$('#parameters').parent().addClass('btn-sm');")
shinyjs::runjs("$('#parameters').parent().parent().parent().parent().parent().addClass('well').addClass('well-sm');")
output$parameters_feedback <- shiny::renderTable({
shiny::req(input$parameters)
params <- vroom::vroom(input$parameters$datapath)
shiny::validate(shiny::need(ncol(params) == 3 & "PARAMETER" %in% names(params) & "VALUE" %in% names(params),
"Incorrect parameters file."))
params$loaded <- FALSE
# Ignored: *_refs, Package version
params <- params[!endsWith(params$PARAMETER, "_refs") & params$PARAMETER != "Package version",]
# checkboxInput: filter_blk_ghost_peaks, compute_pearson_correlation
booleans <- params[params$VALUE %in% c("TRUE", "FALSE"),]$PARAMETER
for (pb in booleans[booleans %in% names(input)]) {
try({
shiny::updateCheckboxInput(session, pb, value = as.logical(params[params$PARAMETER == pb,]$VALUE))
params[params$PARAMETER == pb,]$loaded <- TRUE
})
}
# filter_blk, filter_mz, filter_rsd, filter_rmd -> strings in checkboxGroupInput "filters"
try({
group_filters <- booleans[!(booleans %in% names(input))
& booleans %in% c("filter_blk", "filter_mz", "filter_rsd", "filter_rmd")]
shiny::updateCheckboxGroupInput(session, "filters", selected = group_filters)
for (pf in group_filters) params[params$PARAMETER == pf,]$loaded <- TRUE
})
# selectInput : pearson_p_value
if ("pearson_p_value" %in% params$PARAMETER) {
try({
choices <- c(1, 0.05, 0.01)
diff <- as.numeric(params[params$PARAMETER == "pearson_p_value",]$VALUE) - unlist(choices)
closest <- choices[which.min(abs(diff))]
shiny::updateSelectInput(session, "pearson_p_value", selected = closest)
params[params$PARAMETER == "pearson_p_value",]$loaded <- TRUE
})
}
# filter_rmd_range -> 2 numericInput "filter_rmd_range_min" and "filter_rmd_range_max"
if ("filter_rmd_range" %in% params$PARAMETER) {
try({
str_range <- as.character(params[params$PARAMETER == "filter_rmd_range",]$VALUE)
str_range <- as.numeric(unlist(strsplit(substr(str_range, 2, nchar(str_range)-1), ",")))
shiny::updateNumericInput(session, "filter_rmd_range_min", value = str_range[1])
shiny::updateNumericInput(session, "filter_rmd_range_max", value = str_range[2])
params[params$PARAMETER == "filter_rmd_range",]$loaded <- TRUE
})
}
# selection_criterion -> strings in checkboxGroupInput "selection_criteria" ("intensity", "degree")
if ("selection_criterion" %in% params$PARAMETER) {
if (params[params$PARAMETER == "selection_criterion",]$VALUE %in% c("both", "intensity", "degree")) {
try({
if (params[params$PARAMETER == "selection_criterion",]$VALUE == "both") {
shiny::updateCheckboxGroupInput(session, "selection_criteria", selected = c("intensity", "degree"))
} else {
shiny::updateCheckboxGroupInput(session, "selection_criteria",
selected = c(as.character(params[params$PARAMETER == "selection_criterion",]$VALUE)))
}
params[params$PARAMETER == "selection_criterion",]$loaded <- TRUE
})
}
}
# textInput: compound_levels, biosoc_levels, levels_scores
for (pl in c("compound_levels", "biosoc_levels", "levels_scores")) {
try({
if (pl %in% params$PARAMETER & params[params$PARAMETER == pl,]$VALUE != "[]") {
str_list <- as.character(params[params$PARAMETER == pl,]$VALUE)
str_list <- substr(str_list, 2, nchar(str_list)-1)
shiny::updateTextInput(session, pl, value = str_list)
}
params[params$PARAMETER == pl,]$loaded <- TRUE
})
}
# numericInput
for (pn in c("filter_blk_threshold", "filter_rsd_threshold", "threshold_mz", "threshold_rt",
"pearson_correlation_threshold", "n", "min_score")) {
try({
shiny::updateNumericInput(session, pn, value = as.numeric(params[params$PARAMETER == pn,]$VALUE))
params[params$PARAMETER == pn,]$loaded <- TRUE
})
}
# Feedback for user
params
})
# References
user_ref_generic <- function(ref_file) {
shiny::req(ref_file)
data <- vroom::vroom(ref_file$datapath)
shiny::validate(shiny::need(ncol(data) == 2, "Please upload a reference file with 2 columns (see documentation)."))
return(data)
}
user_pos_adducts_refs <- shiny::reactive({ shiny::req(user_ref_generic(input$user_pos_adducts_refs))
user_ref_generic(input$user_pos_adducts_refs) })
user_neg_adducts_refs <- shiny::reactive({ shiny::req(user_ref_generic(input$user_neg_adducts_refs))
user_ref_generic(input$user_neg_adducts_refs) })
user_pos_neutral_refs <- shiny::reactive({ shiny::req(user_ref_generic(input$user_pos_neutral_refs))
user_ref_generic(input$user_pos_neutral_refs) })
user_neg_neutral_refs <- shiny::reactive({ shiny::req(user_ref_generic(input$user_neg_neutral_refs))
user_ref_generic(input$user_neg_neutral_refs) })
preview_ref <- function(ref_data) rbind(utils::head(ref_data, 2),
stats::setNames(data.frame(a = "...", b = "..."), names(ref_data)))
output$user_pos_adducts_refs <- shiny::renderTable(preview_ref(user_pos_adducts_refs()))
output$user_neg_adducts_refs <- shiny::renderTable(preview_ref(user_neg_adducts_refs()))
output$user_pos_neutral_refs <- shiny::renderTable(preview_ref(user_pos_neutral_refs()))
output$user_neg_neutral_refs <- shiny::renderTable(preview_ref(user_neg_neutral_refs()))
# clean_msdial_data
clean_params <- shiny::eventReactive(input$button_clean, {
check_ref <- function(ref) {
tryCatch({ shiny::validate(shiny::need(ref, "not found")) ; return(ref) },
error = function(e) {})
}
get_ref <- function(ref) if(is.null(check_ref(ref))) NA else as.data.frame(ref)
list(
filter_blk = "filter_blk" %in% input$filters,
filter_blk_threshold = input$filter_blk_threshold,
filter_blk_ghost_peaks = input$filter_blk_ghost_peaks,
filter_mz = "filter_mz" %in% input$filters,
filter_rsd = "filter_rsd" %in% input$filters,
filter_rsd_threshold = input$filter_rsd_threshold,
filter_rmd = "filter_rmd" %in% input$filters,
filter_rmd_range = c(input$filter_rmd_range_min, input$filter_rmd_range_max),
threshold_mz = input$threshold_mz,
threshold_rt = input$threshold_rt,
compute_pearson_correlation = input$compute_pearson_correlation,
pearson_correlation_threshold = input$pearson_correlation_threshold,
pearson_p_value = input$pearson_p_value,
user_pos_adducts_refs = get_ref(user_pos_adducts_refs()),
user_neg_adducts_refs = get_ref(user_neg_adducts_refs()),
user_pos_neutral_refs = get_ref(user_pos_neutral_refs()),
user_neg_neutral_refs = get_ref(user_neg_neutral_refs())
)
})
output$clean <- shiny::renderPrint({
shiny::req(input$button_clean > 0)
shiny::req(exists("analysis_directory", envir = mscleanrCache))
shiny::req(clean_params())
waiter <- fun_waiter("Cleaning data...")
waiter$show()
tryCatch(
do.call(clean_msdial_data, clean_params()),
finally = waiter$hide()
)
})
# keep_top_peaks
keep_params <- shiny::eventReactive(input$button_keep, {
shiny::req("degree" %in% input$selection_criteria | "intensity" %in% input$selection_criteria)
if ("degree" %in% input$selection_criteria) {
if ("intensity" %in% input$selection_criteria) mode <- "both"
else mode <- "degree"
} else mode <- "intensity"
list(selection_criterion = mode,
n = input$n,
export_filtered_peaks = input$export_filtered_peaks)
})
output$keep <- shiny::renderPrint({
shiny::req(input$button_keep > 0)
shiny::req(exists("analysis_directory", envir = mscleanrCache))
shiny::req(keep_params())
waiter <- fun_waiter("Keeping only selected peaks...")
waiter$show()
tryCatch(
do.call(keep_top_peaks, keep_params()),
finally = waiter$hide()
)
})
# launch_msfinder_annotation
clean <- function(l) {
tmp <- unlist(lapply(strsplit(l, ",", fixed = TRUE), function(x) trimws(x)))
tmp <- tmp[tmp != ""]
if (length(tmp) == 0) return(NULL) else return(unique(tmp))
}
compounds <- shiny::reactive(clean(input$compound_levels))
biosoc <- shiny::reactive({
tmp <- clean(input$biosoc_levels)
if (is.null(tmp)) c("generic") else tmp
})
scores <- shiny::reactive({
tmp_scores <- lapply(clean(input$levels_scores), strsplit, split = ":", fixed = TRUE)
if (length(tmp_scores) == 0) scores <- NULL
else {
scores <- list()
for (x in tmp_scores) scores[x[[1]][1]] <- ifelse(is.na(x[[1]][2]) | x[[1]][2] == "",
1,
as.numeric(x[[1]][2]))
}
scores
})
output$compound_levels <- shiny::renderTable({
shiny::req(compounds())
data.frame("Rank" = c(1:length(compounds())), "Compound level" = compounds())
})
output$biosoc_levels <- shiny::renderTable(
data.frame("Rank" = c(1:length(unique(c(biosoc(), "generic")))),
"Biosource level" = unique(c(biosoc(), "generic")))
)
output$levels_scores <- shiny::renderTable(
data.frame(Level = names(scores()), Multiplier = unlist(scores(), use.names = FALSE))
)
launch_params <- shiny::eventReactive(input$button_launch, {
list(compound_levels = compounds(),
biosoc_levels = biosoc(),
levels_scores = scores(),
score_only = input$score_only,
other_peaks_warning = input$other_peaks_warning)
})
output$launch <- shiny::renderPrint({
shiny::req(input$button_launch > 0)
shiny::req(exists("analysis_directory", envir = mscleanrCache))
shiny::req(launch_params())
waiter <- fun_waiter("Annotating peaks with MS-FINDER data...")
waiter$show()
tryCatch(
do.call(launch_msfinder_annotation, launch_params()),
finally = waiter$hide()
)
})
# convert_csv_to_msp
convert_params <- shiny::eventReactive(input$button_convert, list(all = input$all,
min_score = input$min_score))
output$convert <- shiny::renderPrint({
shiny::req(input$button_convert > 0)
shiny::req(exists("analysis_directory", envir = mscleanrCache))
shiny::req(convert_params())
waiter <- fun_waiter("Converting peaks to MSP...")
waiter$show()
tryCatch(
do.call(convert_csv_to_msp, convert_params()),
finally = waiter$hide()
)
})
# CAD/PDA
# cad_clean_params <- shiny::eventReactive(input$button_cad_clean, {
# list(mode = input$mode,
# data_start_row = input$data_start_row,
# plot_spectra = input$plot_spectra,
# plot_ROI = input$plot_ROI)
# })
#
# output$cad_clean <- shiny::renderPrint({
# shiny::req(input$button_cad_clean > 0)
# shiny::req(exists("analysis_directory", envir = mscleanrCache))
# shiny::req(cad_clean_params())
# waiter <- fun_waiter(paste("Cleaning", cad_clean_params()$mode, "peaks..."))
# waiter$show()
# tryCatch(
# do.call(clean_cad_pda_peaks, cad_clean_params()),
# finally = waiter$hide()
# )
# })
#
# cad_merge_params <- shiny::eventReactive(input$button_cad_merge, {
# list(mode = input$mode,
# delta_rt = input$delta_rt,
# threshold_rt = input$cad_threshold_rt,
# pearson_correlation_threshold = input$cad_pearson_correlation_threshold,
# pearson_p_value = input$cad_pearson_p_value)
# })
#
# output$cad_merge <- shiny::renderPrint({
# shiny::req(input$button_cad_merge > 0)
# shiny::req(exists("analysis_directory", envir = mscleanrCache))
# shiny::req(cad_merge_params())
# waiter <- fun_waiter(paste("Merging", cad_merge_params()$mode, "and MS peaks..."))
# waiter$show()
# tryCatch(
# do.call(launch_cad_fusion, cad_merge_params()),
# finally = waiter$hide()
# )
# })
# Datasets
output$adducts_pos <- shiny::renderDataTable(mass_adducts_pos, escape = FALSE, options = list(pageLength = 10))
output$adducts_neg <- shiny::renderDataTable(mass_adducts_neg, escape = FALSE, options = list(pageLength = 10))
output$nl_pos <- shiny::renderDataTable(mass_neutral_loss_pos, escape = FALSE, options = list(pageLength = 10))
output$nl_neg <- shiny::renderDataTable(mass_neutral_loss_neg, escape = FALSE, options = list(pageLength = 10))
output$isotopes <- shiny::renderDataTable(mass_isotopes, escape = FALSE, options = list(pageLength = 10))
dl <- function(data, filename) shiny::downloadHandler(filename = filename,
content = function(file) vroom::vroom_write(data, file, delim = ","))
output$dl_adducts_pos <- dl(mass_adducts_pos, "positive_adducts.csv")
output$dl_adducts_neg <- dl(mass_adducts_neg, "negative_adducts.csv")
output$dl_nl_pos <- dl(mass_neutral_loss_pos, "positive_neutral_losses.csv")
output$dl_nl_neg <- dl(mass_neutral_loss_neg, "negative_neutral_losses.csv")
shiny::onStop(function() { assign("shiny_running", FALSE, envir = mscleanrCache) })
}
shiny::shinyApp(ui = ui,
server = server,
onStart = function() { assign("shiny_running", TRUE, envir = mscleanrCache) })
}
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.