R/batman_helpers.R

In AlpsNMR: Automated spectraL Processing System for NMR

#' Batman helpers
#'
#' @return These are helper functions to make Batman tests easier
#'
#' @param bopts Batman options
#' @param batman_dir Batman input directorye
#' @param filename Filename to use, inside `batman_dir`
#' @param metabolite_names A character vector of the metabolite names to consider
#' @param nmr_dataset An [nmr_dataset_1D] object
#' @param multiplet_table A data frame, like the [hmdb] dataset
#'
#' @family batman functions
#' @name nmr_batman
#'
NULL

#' The Human Metabolome DataBase multiplet table
#'
#' @name hmdb
#' @docType data
#' @references \url{https://hmdb.ca/}
#' @keywords data
NULL


#' Batman Options helper
#'
#' @param ppmRange Range of ppm to process
#' @param specNo Index of spectra to process
#' @param paraProc Number of cores to use
#' @param negThresh Truncation threshold for negative intensities
#' @param scaleFac Divide each spectrum by this number
#' @param downSamp Decimate each spectrum by this factor
#' @param hiresFlag Keep High Resolution deconvolved spectra
#' @param randSeed A random seed
#' @param nItBurnin Number of burn-in iterations
#' @param nItPostBurnin Number of iterations after burn-in
#' @param multFile Multiplet file (integer)
#' @param thinning Save MCMC state every thinning iterations
#' @param cfeFlag Same concentration for all spectra (fixed effect)
#' @param nItRerun Number of iterations for a batman rerun
#' @param startTemp Start temperature
#' @param specFreq NMR Spectrometer frequency
#' @param a Shape parameter for the gamma distribution (used for lambda, the precision)
#' @param b Rate distribution parameter for the gamma distribution (used for lambda, the precision)
#' @param muMean Peak width mean in ln(Hz)
#' @param muVar Peak width variance in ln(Hz)
#' @param muVar_prop Peak width proposed variance in ln(Hz)
#' @param nuMVar Peak width metabolite variance in ln(Hz)
#' @param nuMVarProp Peak width metabolite proposed variance in ln(Hz)
#' @param tauMean mean of the prior on tau
#' @param tauPrec inverse of variance of prior on tau
#' @param rdelta Truncation of the prior on peak shift (ppm)
#' @param csFlag Specify chemical shift for each multiplet in each spectrum? (chemShiftperSpectra.csv file)
#'
#' @return A batman_options object with the Batman Options
#' @family batman functions
#' @export
#' @examples
#' bopts <- nmr_batman_options()
nmr_batman_options <- function(ppmRange = matrix(
    c(3.0, 3.1,
      3.6, 3.7,
      3.9, 4.0,
      4.0, 4.1,
      6.95, 7.05,
      7.6, 7.7,
      7.8, 7.9),
    ncol = 2,
    byrow = TRUE
),
specNo = "1",
paraProc = 4L,
negThresh = -0.5,
scaleFac = 1000000,
downSamp = 1,
hiresFlag = 1,
randSeed = 100025L,
nItBurnin = 200L,
nItPostBurnin = 5000L,
multFile = 2L,
thinning = 50L,
cfeFlag = 0,
nItRerun = 5000L,
startTemp = 1000,
specFreq = 600,
a = 0.00001,
b = 0.000000001,
muMean = 1.1,
muVar = 0.2,
muVar_prop = 0.002,
nuMVar = 0.0025,
nuMVarProp = 0.1,
tauMean = -0.05,
tauPrec = 2,
rdelta = 0.02,
csFlag = 0) {
    # ppmRange:
    ppmRange <- as.matrix(ppmRange)
    stopifnot(ncol(ppmRange) == 2)
    stopifnot(!anyNA(ppmRange))
    
    out <- list(
        ppmRange = ppmRange,
        specNo = specNo,
        paraProc = paraProc,
        negThresh = negThresh,
        scaleFac = scaleFac,
        downSamp = downSamp,
        hiresFlag = hiresFlag,
        randSeed = randSeed,
        nItBurnin = nItBurnin,
        nItPostBurnin = nItPostBurnin,
        multFile = multFile,
        thinning = thinning,
        cfeFlag = cfeFlag,
        nItRerun = nItRerun,
        startTemp = startTemp,
        specFreq = specFreq,
        a = a,
        b = b,
        muMean = muMean,
        muVar = muVar,
        muVar_prop = muVar_prop,
        nuMVar = nuMVar,
        nuMVarProp = nuMVarProp,
        tauMean = tauMean,
        tauPrec = tauPrec,
        rdelta = rdelta,
        csFlag = csFlag
    )
    class(out) <- "batman_options"
    out
}

#' @rdname nmr_batman
#' @export
#' @examples
#' bopts <- nmr_batman_options()
#' # nmr_batman_write_options(bopts)
#' 
nmr_batman_write_options <- function(bopts,
                                     batman_dir = "BatmanInput",
                                     filename = "batmanOptions.txt") {
    # ppmRange:
    full_filename <-
        batman_get_full_filename(batman_dir, filename)
    bopts2 <- bopts
    ppmRange <- as.matrix(bopts2$ppmRange)
    stopifnot(ncol(ppmRange) == 2)
    stopifnot(!anyNA(ppmRange))
    ppmRange <-
        paste0("(", apply(ppmRange, 1, function(x)
            paste(x, collapse = ",")), ")", collapse = " ")
    bopts2$ppmRange <- ppmRange
    # specNo
    specNo <- as.numeric(bopts2$specNo)
    specNo <- paste0(specNo, collapse = ",")
    bopts2$specNo <- specNo
    
    bopts_txt <- glue::glue_data(
        bopts2,
        "%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Options file for BATMAN %%%%%%%%%%%%%%%%%%%%%%%%%",
        "%% There should be no empty line in the file, if line                               %%",
        "%% is not used, please comment the start of the line with %.                        %%",
        "%% Please do not comment or remove any of the parameter input lines.                %%",
        "%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%",
        "%%",
        "%% General parameters:",
        "%% ppm ranges for analysis: Please put each ppm range in a pair of parentheses in the same line, ",
        "%% separate start and end ppm values with a comma, and ",
        "%% separate each set of ppm range with space, e.g. (-0.05,0.05) (1.44,1.52)",
        "ppmRange - ppm ranges for analysis: {ppmRange}",
        "specNo - Ranges of spectra number to be included (e.g. 1,3-4 etc.): {specNo}",
        "paraProc - No of parallel processes (multicores) (only 1 core will be used for single spectrum): {paraProc}",
        "negThresh - Truncation threshold for negative intensities: {negThresh}",
        "scaleFac - Intensity scale factor: {scaleFac}",
        "downSamp - Down sampling factor: {downSamp}",
        "hiresFlag - Save metabolite fit at resolution of original spectrum? (Yes - 1 / No - 0): {hiresFlag}",
        "randSeed - Random number seed: {randSeed}",
        "nItBurnin - Number of burn-in iterations: {nItBurnin}",
        "nItPostBurnin - Number of post-burn-in iterations: {nItPostBurnin}",
        "multFile - Choose template of multiplets file from options below: {multFile}",
        "%% 1, The default template of multiplets in multi_data.csv file,",
        "%% 2, The user input template of multiplets in multi_data_user.csv file,",
        "%% 3, Both the default and user input template of multiplets files.",
        "%%",
        "thinning - Save MCMC state in every ? iterations: {thinning}",
        "cfeFlag - Same concentration for all spectra (fixed effect)? (Yes - 1 / No - 0): {cfeFlag}",
        "nItRerun - Number of iterations for batmanrerun: {nItRerun}",
        "startTemp - Start temperature: {startTemp}",
        "specFreq - Spectrometer frequency (MHz): {specFreq}",
        "%% ",
        "%% Uncatalogued (wavelet) component: ",
        "%% Hyper parameters for the global precision priors, lambda is inversely proportional to error variance. ",
        "%% lambda ~ Gamma(a,b/2) on wavelet coefficients:",
        "a - Gamma-distributed with shape a: {a}",
        "b - Gamma-distributed with scale b: {b}",
        "%% ",
        "%% Catalogued metabolite component: gamma is the peak width (full width at half max) of metabolite m. ",
        "%% The model for gamma is ln(gamma)= mu + nu_m, where mu is the spectrum-wide average log-peak width ",
        "%% and nu_m is a random effect for each metabolite deviating from mu.",
        "muMean - Mean of prior on global peak width (mu) in ln(Hz): {muMean} ",
        "muVar - Variance of prior on global peak width (mu) in ln(Hz): {muVar}",
        "muVar_prop - Variance of proposal distribution for mu in ln(Hz): {muVar_prop}",
        "%% The mean of each prior on nu_m is 0. ",
        "%% Set variance of prior on peak width offset (nu_m) to 0 to turn off the random effect on peak width: ",
        "nuMVar - Variance of prior on peak width offset (nu_m) in ln(Hz): {nuMVar}",
        "nuMVarProp - Variance of proposal distribution for nu_m in ln(Hz): {nuMVarProp}",
        "%%",
        "%% Wavelet truncation: tau is a vector of truncation limits, which sets a lower bound on the wavelets",
        "tauMean - mean of the prior on tau: {tauMean}",
        "tauPrec - inverse of variance of prior on tau: {tauPrec}",
        "%%",
        "%% Global prior on peak shift (truncated Gaussian)",
        "%% Note: individual priors for each multiplet can be changed in the multi_data(_user).csv file",
        "rdelta - Truncation of the prior on peak shift (ppm): {rdelta}",
        "csFlag - Specify chemical shift for each multiplet in each spectrum? (chemShiftperSpectra.csv file) (Yes - 1 / No - 0): {csFlag}",
        .sep = "\n"
    )
    writeLines(text = bopts_txt, con = full_filename)
    bopts
}


batman_get_full_filename <- function(batman_dir, filename) {
    if (!dir.exists(batman_dir)) {
        dir.create(batman_dir, recursive = TRUE)
    }
    full_filename <- file.path(batman_dir, filename)
    if (file.exists(full_filename)) {
        stop("File ",
             full_filename,
             " already exists. Remove it and try again.")
    }
    full_filename
}

#' @rdname nmr_batman
#' @export
#' @examples
#' dir_to_demo_dataset <- system.file("dataset-demo", package = "AlpsNMR")
#' dataset <- nmr_read_samples_dir(dir_to_demo_dataset)
#' dataset_1D <- nmr_interpolate_1D(dataset, axis = c(min = -0.5, max = 10, by = 2.3E-4))
#' #nmr_batman_export_dataset(dataset_1D)
#'
nmr_batman_export_dataset <- function(nmr_dataset,
                                      batman_dir = "BatmanInput",
                                      filename = "NMRdata.txt") {
    full_filename <- batman_get_full_filename(batman_dir, filename)
    nmrdata <- t(nmr_data(nmr_dataset))
    colnames(nmrdata) <-
        paste0("NMRExperiment_", colnames(nmrdata))
    nmrdata <-
        cbind(ppm = as.numeric(rownames(nmrdata)), nmrdata)
    utils::write.table(nmrdata,
                       full_filename,
                       row.names = FALSE,
                       sep = "\t")
}

#' @rdname nmr_batman
#' @export
#' @examples
#' message("Use of multi_data_user_hmdb")
#' #multi_data_user_hmdb <- nmr_batman_multi_data_user_hmdb()
nmr_batman_multi_data_user_hmdb <- function(batman_dir = "BatmanInput",
                                            filename = "multi_data_user.csv") {
    hmdb <- NULL
    utils::data("hmdb", package = "AlpsNMR", envir = environment())
    hmdb <- nmr_batman_multi_data_user(multiplet_table = hmdb,
                                       batman_dir = batman_dir,
                                       filename = filename)
    hmdb
}

#' @rdname nmr_batman
#' @export
#' @examples
#' hmdb <- NULL
#' #utils::data("hmdb", package = "AlpsNMR", envir = environment())
#' #hmdb <- nmr_batman_multi_data_user(hmbd)
#'
nmr_batman_multi_data_user <- function(multiplet_table,
                                       batman_dir = "BatmanInput",
                                       filename = "multi_data_user.csv") {
    full_filename <- batman_get_full_filename(batman_dir, filename)
    
    # column default values: Create them if they don't exist
    col_def <-
        data.frame(
            name = c(
                "overwrite_pos",
                "overwrite_truncation",
                "Include_multiplet"
            ),
            value = c(-50, -50, 1),
            stringsAsFactors = FALSE
        )
    
    for (i in seq_len(nrow(col_def))) {
        col_name <- col_def[["name"]][i]
        col_val <- col_def[["value"]][i]
        if (!col_name %in% colnames(multiplet_table)) {
            multiplet_table[[col_name]] <- col_val
        }
    }
    
    # Columns to be written, as needed by Batman
    cols <- c(
        "Metabolite",
        "pos_in_ppm",
        "couple_code",
        "J_constant",
        "relative_intensity",
        "overwrite_pos",
        "overwrite_truncation",
        "Include_multiplet"
    )
    
    multiplet_table[["J_constant"]][is.na(multiplet_table[["J_constant"]])] <-
        0 # batman does not accept NA...
    multiplet_table[["Metabolite"]] <-
        gsub(pattern = ",",
             replacement = "_",
             multiplet_table[["Metabolite"]])
    utils::write.csv(multiplet_table[, cols], full_filename, row.names = FALSE)
    multiplet_table
}

#' @rdname nmr_batman
#' @export
#' @examples
#' metabolite_names <- c("alanine", "glucose")
#' #metabolite_names <- nmr_batman_metabolites_list(metabolite_names)
#'
nmr_batman_metabolites_list <- function(metabolite_names,
                                        batman_dir = "BatmanInput",
                                        filename = "metabolitesList.csv") {
    full_filename <- batman_get_full_filename(batman_dir, filename)
    write(unique(metabolite_names), file = full_filename)
    metabolite_names
}