R/process_plate.R
In ec363/fpcountr: Fluorescent Protein Calibration For Plate Readers
Defines functions
process_plate
Documented in
process_plate
#' Process experimental data (normalise, correct and calibrate)
#'
#' Processes experimental (fluorescence and absorbance) data from microplate
#' readers. Expects 'parsed' data that is tidy and attached to appropriate
#' metadata. Numerous arguments offer options to customise how and whether
#' normalisation, correction and/or calibration is carried out. Data can be
#' timecourse (kinetic) data or single timepoint (endpoint) data, and can
#' contain fluorescence data only or include absorbance (OD) data to enumerate
#' cell density. Originally based on `flopr::process_plate()` but with numerous
#' changes. First, the function normalises optical density readings (if present)
#' to media/buffer wells, and fluorescence readings, either to media/buffer
#' wells or to cellular autofluorescence. Fluorescence channels must be
#' specified in `flu_channels`, the blank wells in `blank_wells`, and the
#' reading to be used for optical density in `od_name`. The cellular
#' autofluorescence model can be set with `af_model` on the negative
#' (`neg_well`) wells, or set to `NULL`. Second, if `do_quench_correction` is
#' TRUE, the 'quenching' or attenuation of fluorescence measurements by cells is
#' corrected for according to the cell density measured. This requires the
#' specification of `od_type` as "OD600" or "OD700". Third, if `do_calibrate` is
#' TRUE, normalised values for both fluorescence and optical density are
#' converted into meaningful units. Here, the paths to conversion factor CSV
#' files must be provided as `od_coeffs_csv` for OD and `fluor_coeffs_csv` for
#' fluorescence. For fluorescence measurements, each calibration must be
#' specified by `flu_slugs` to represent the FP used, `flu_gains` to provide the
#' gain, and `flu_labels` to specify how the relevant plots should be labelled,
#' (e.g. `flu_slugs = c("mcherry", "mtagbfp2")`, `flu_gains = c(60,80)`,
#' `flu_labels = c("RFP, BFP")`).
#'
#' @param data_csv path to a CSV file containing parsed plate reader data
#' @param blank_well the well coordinates of one or more media blanks. Defaults
#' to "A1".
#' @param timecourse logical. Is the data timecourse/kinetic data and does it
#' include a variable called 'time'?
#' @param od_name the column name for the optical density data. Defaults to
#' "OD". If no OD measurements were taken, use `NULL`.
#' @param flu_channels the column names for the fluorescence data. Defaults to
#' "green1green2".
#' @param flu_channels_rename if specified, what to change the flu_channels
#' column names to, before computing normalisations and calibrations. Can be
#' useful if the channel_name in the calibration file is named differently
#' from the columns in the data file. Needs to be same length as flu_channels,
#' if not all require changing, specify them anyway to allow positional
#' replacement (first element in flu_channels_rename replaces first in
#' flu_channels, etc.). Defaults to NULL.
#' @param af_model model used to fit negative control autofluorescence. For now
#' these include "polynomial", "inverse_poly", "exponential", "spline" and
#' "loess". If set to NULL, no model is used, and fluorescence is normalised
#' akin to OD: by subtracting the value for the blanks. Defaults to "spline".
#' @param neg_well the well coordinates of a non-fluorescent control. Defaults
#' to "A2".
#' @param do_quench_correction logical. Should function correct for anticipated
#' quenching of fluorescence, depending on the cell density?
#' @param od_type Which OD-type was used? Required for quench correction.
#' "OD600" or "OD700".
#' @param do_calibrate logical. Should function convert OD and fluorescence data
#' to calibrated units? Defaults to FALSE.
#' @param instr character string to represent instrument. If do_calibrate =
#' TRUE, used for filtering `od_coeffs_csv` and `fluor_coeffs_csv` files for
#' conversion factors of the relevant instrument.
#' @param flu_slugs character array representing fluorescent proteins (format =
#' FPbase slug). If do_calibrate = TRUE, used for filtering `fluor_coeffs_csv`
#' for conversion factors of the relevant FP.
#' @param flu_gains numeric array representing gains of each fluorescent
#' channel. If do_calibrate = TRUE, used for filtering `fluor_coeffs_csv` for
#' conversion factors of the relevant gain.
#' @param flu_labels If do_calibrate = TRUE, the column names to be given to
#' each calibration. May be identical to flu_slug or flu_channel, but
#' recommended is to make it obvious which FP is being calibrated, e.g.
#' "mCherry", as channel names may be non-specifically named e.g. "red1red1".
#' Needs to be same length as flu_slugs and flu_gains.
#' @param od_coeffs_csv if do_calibrate = TRUE, path of the CSV file containing
#' conversion factors for optical density
#' @param fluor_coeffs_csv if do_calibrate = TRUE, path of the CSV file
#' containing predicted conversion factors for the fluorescent channels
#' @param outfolder path to folder where output files should be saved. Defaults
#' to current working directory.
#' @param csv_only logical. Saves only CSV files as outputs when `TRUE`.
#' Defaults to `FALSE`.
#'
#' @return a data.frame with columns for raw plate reader data, normalised data
#' and, if do_calibrate = TRUE, calibrated OD and FP data
#' @export
#' @importFrom rlang .data
#'
#' @examples
#' \dontrun{
#' processed_data <- process_plate(
#' data_csv = "mcherry_parsed.csv",
#' blank_well = c("A11"),
#' od_name = "OD600", flu_channels = c("red1"), flu_channels_rename = c("red1red1"),
#' af_model = NULL,
#' do_quench_correction = TRUE, od_type = "OD700",
#' do_calibrate = TRUE, instr = "spark1",
#' flu_slugs = c("mcherry"), flu_gains = c(80), flu_labels = c("mcherry"),
#' od_coeffs_csv = "od_coeffs.csv", fluor_coeffs_csv = "flu_coeffs.csv",
#' outfolder = file.path("data_processed")
#' )
#' }
process_plate <- function(
data_csv, blank_well = "A1",
# timecourse
timecourse = TRUE,
# od
od_name = "OD",
# fluorescence
flu_channels = c("green1green2"), # column names in data
flu_channels_rename = NULL, # if not NULL, for every flu_channels specify new name
# rename data columns to make sure they match conversion factor table entries
# autofluorescence model
af_model = "spline", # options: NULL, "spline", "loess"
neg_well = "A2", # if af_model = NULL this can be hashed
# cell quench correction
do_quench_correction = FALSE,
od_type, # "OD600" or "OD700"
# calibrations
do_calibrate = FALSE,
instr, # instrument name for the calibration(s)
# flu_channels above specifies channel(s) for the calibration(s)
flu_slugs = c(), # FP name in FPbase's slug format for the calibration(s)
# OR whatever matches the format used in the conversion factor tables
flu_gains = c(), # gain for calibration(s)
# each of the above need to be specified for every flu_channels being calibrated
flu_labels = c(), # how to display the FP/channel combination in plots and file names
# important if slug is unhelpful or multiple gains used for same FP
# conversion factors
od_coeffs_csv, # microsphere conversion factors for od600, od700
fluor_coeffs_csv, # FP conversion factors
outfolder = ".",
csv_only = FALSE
) {
# Get parsed data --------------------------------------------------------------------------------------------------
pr_data <- utils::read.csv(data_csv, check.names = FALSE)
pr_data
# Rename flu channels if requested ---------------------------------------------------------------------------------
if(!is.null(flu_channels_rename)){
## Check there are replacements for each flu channel
if(!isTRUE(all.equal(length(flu_channels), length(flu_channels_rename), check.attributes = FALSE))){
message("Error: number of flu_channels must match number of flu_channels_rename.
To remove requirement for renaming, remove flu_channels_rename or set it to NULL.")
return()
}
## Rename columns
for (flu_name_idx in seq_len(length(flu_channels))){
names(pr_data)
col_idx <- which(names(pr_data) == flu_channels[flu_name_idx])
names(pr_data)[col_idx] <- flu_channels_rename[flu_name_idx]
names(pr_data)
}
# Rest of script thinks flu_channels are the final column names needed.
# So flu_channels needs to be overwritten with flu_channels_rename to prevent downstream errors
flu_channels <- flu_channels_rename
}
# Location for saved plots -----------------------------------------------------------------------------------------
# make folder if it doesn't exist already
ifelse(test = !dir.exists(file.path(outfolder)), yes = dir.create(file.path(outfolder)), no = FALSE)
# OD processing sections --------------------------------------------------------------------
if(is.null(od_name)){
# skip OD processing...
# rename data
od_norm_pr_data <- pr_data # otherwise done at od_norm() step # NB. this won't have normalised_OD or normalised_OD_cm1 columns
} else {
# proceed with OD processing...
# Normalise OD (with external od_norm function) --------------------------------------------------------------------
od_norm_pr_data <- fpcountr::od_norm(pr_data, blank_well, od_name, timecourse = timecourse)
# adds one extra column - normalised_OD. for each timepoint, normalises OD to mean of blank wells
# Plot OD data in plate format
# plot
if(isFALSE(timecourse)){
# heatmap1 - raw OD
max_value <- max(od_norm_pr_data[[od_name]], na.rm = TRUE)
plt_od <- ggplot2::ggplot(data = od_norm_pr_data,
# ggplot2::aes(x = row, y = column, fill = .data[[od_name]])) +
ggplot2::aes(x = .data$column, y = row, fill = .data[[od_name]])) +
ggplot2::geom_tile() +
ggplot2::scale_x_discrete("", position = "top", # put axis labels at top
limits = factor(unique(od_norm_pr_data$column))) + # put first element at top # factor to make it discrete rather than continuous
ggplot2::scale_y_discrete("", limits = rev(unique(od_norm_pr_data$row))) + # put first element at top
viridis::scale_fill_viridis("raw OD",
discrete = FALSE, limits = c(0, max_value),
alpha = 0.4,
na.value = "white") +
ggplot2::geom_text(ggplot2::aes(label = round(.data[[od_name]], 2)), na.rm = TRUE, size = 2.5) +
# ggplot2::coord_fixed(ratio = 1) +
ggplot2::theme_bw() + # base_size = 8
ggplot2::theme(
aspect.ratio = 8/12,
panel.grid = ggplot2::element_blank()
)
plt_od
plotname <- "OD_1a_raw.pdf"
if(isFALSE(csv_only)){
ggplot2::ggsave(file.path(outfolder, plotname),
plot = plt_od,
height = 16, width = 24, units = "cm")
}
# normalised OD
max_value <- max(od_norm_pr_data$normalised_OD, na.rm = TRUE)
plt_od <- ggplot2::ggplot(data = od_norm_pr_data,
ggplot2::aes(x = .data$column, y = row, fill = .data$normalised_OD)) +
ggplot2::geom_tile() +
ggplot2::scale_x_discrete("", position = "top", limits = factor(unique(od_norm_pr_data$column))) +
ggplot2::scale_y_discrete("", limits = rev(unique(od_norm_pr_data$row))) +
viridis::scale_fill_viridis("normalised OD",
discrete = FALSE, limits = c(0, max_value),
alpha = 0.4,
na.value = "white") +
ggplot2::geom_text(ggplot2::aes(label = round(.data$normalised_OD, 2)), na.rm = TRUE, size = 2.5) +
ggplot2::theme_bw() + # base_size = 8
ggplot2::theme(
aspect.ratio = 8/12,
panel.grid = ggplot2::element_blank()
)
plt_od
plotname <- "OD_1b_normalised.pdf"
if(isFALSE(csv_only)){
ggplot2::ggsave(file.path(outfolder, plotname),
plot = plt_od,
height = 16, width = 24, units = "cm")
}
} else if(isTRUE(timecourse)){
# plot with caption
plt_od <- ggplot2::ggplot(od_norm_pr_data) +
ggplot2::geom_line(ggplot2::aes(x = .data$time, y = .data[[od_name]]),
colour = "black", linewidth = 0.5) +
ggplot2::geom_line(ggplot2::aes(x = .data$time, y = .data$normalised_OD),
colour = "red", linewidth = 0.5) +
ggplot2::scale_x_continuous("time") +
ggplot2::labs(caption = "black: raw, red: normalised") +
ggplot2::scale_colour_discrete("") +
ggplot2::facet_grid(row~column) +
ggplot2::theme_bw(base_size = 8) +
ggplot2::theme(
aspect.ratio = 1,
legend.position = "none",
axis.text.x = ggplot2::element_text(angle = 90, vjust = 0.5),
panel.grid.minor = ggplot2::element_blank()
)
plt_od
plotname <- "OD_1_raw_normalised.pdf"
if(isFALSE(csv_only)){
ggplot2::ggsave(file.path(outfolder, plotname),
plot = plt_od,
height = 16, width = 24, units = "cm")
}
}
# Convert to OD (cm-1) --------------------------------------------------------------------
if("volume" %in% names(od_norm_pr_data)){ # only run if data contains the column 'volume'
od_norm_pr_data <- od_norm_pr_data %>%
dplyr::mutate(pathlength = fpcountr::get_pathlength(.data$volume)) %>% # get (estimated) pathlength from volume
dplyr::mutate(normalised_OD_cm1 = .data$normalised_OD/.data$pathlength) # get OD (cm-1)
od_norm_pr_data[1,]
# Plot in plate format
if(isFALSE(timecourse)){
# heatmap1 - raw OD
max_value <- max(od_norm_pr_data$normalised_OD_cm1, na.rm = TRUE)
plt_od <- ggplot2::ggplot(data = od_norm_pr_data,
ggplot2::aes(x = .data$column, y = row, fill = .data$normalised_OD_cm1)) +
ggplot2::geom_tile() +
ggplot2::scale_x_discrete("", position = "top", limits = factor(unique(od_norm_pr_data$column))) +
ggplot2::scale_y_discrete("", limits = rev(unique(od_norm_pr_data$row))) +
viridis::scale_fill_viridis(paste0("normalised ", od_name, " (cm-1)"),
discrete = FALSE, limits = c(0, max_value),
alpha = 0.4,
na.value = "white") +
ggplot2::geom_text(ggplot2::aes(label = round(.data$normalised_OD_cm1, 2)), na.rm = TRUE, size = 2.5) +
ggplot2::theme_bw() + # base_size = 8
ggplot2::theme(
aspect.ratio = 8/12,
panel.grid = ggplot2::element_blank()
)
plt_od
} else if(isTRUE(timecourse)){
plt_od <- ggplot2::ggplot(od_norm_pr_data) +
ggplot2::geom_line(ggplot2::aes(x = .data$time, y = .data$normalised_OD_cm1,
colour = "normalised_cm1"), linewidth = 0.5) +
ggplot2::scale_x_continuous("time") +
ggplot2::scale_y_continuous(name = paste0("normalised ", od_name, " (cm-1)")) + # "normalised OD (cm-1)"
ggplot2::labs(caption = "") +
ggplot2::scale_colour_discrete("") +
ggplot2::facet_grid(row~column) +
ggplot2::theme_bw(base_size = 8) +
ggplot2::theme(
aspect.ratio = 1,
legend.position = "none",
axis.text.x = ggplot2::element_text(angle = 90, vjust = 0.5),
panel.grid.minor = ggplot2::element_blank()
)
plt_od
}
plotname <- "OD_2_pathlength-normalised.pdf"
if(isFALSE(csv_only)){
ggplot2::ggsave(file.path(outfolder, plotname),
plot = plt_od,
height = 16, width = 24, units = "cm")
}
}
} # od processing sections
# Normalise fluorescence data --------------------------------------------------------------------------------------------
flu_norm_pr_data <- od_norm_pr_data
## for each fluorescence channel...
for (flu_idx in seq_len(length(flu_channels))) {
## for each fluorescent protein, run fluorescent normalisation function..
# Force od_name = NULL data to be af_model = NULL (if by accident it isn't)
if(is.null(od_name) & !is.null(af_model)){
message("Autofluorescence model function cannot be run without OD data. Setting af_model to NULL and normalising fluorescence to blank_wells...")
af_model <- NULL
}
# Force timecourse = FALSE data to be af_model = NULL (if by accident it isn't)
if(isFALSE(timecourse) & !is.null(af_model)){
message("Autofluorescence model function cannot be run without timecourse data. Setting af_model to NULL and normalising fluorescence to blank_wells...")
af_model <- NULL
}
flu_norm_pr_data <- fpcountr::flu_norm(pr_data = flu_norm_pr_data, neg_well = neg_well, blank_well = blank_well,
flu_name = flu_channels[flu_idx], af_model = af_model, data_csv = data_csv,
timecourse = timecourse, outfolder = outfolder)
# adds one column per FP of (eg) normalised_[GFP] to the table
# Plot in plate format
if(isFALSE(timecourse)){
# heatmap1 - raw fluor
max_value <- max(flu_norm_pr_data[[flu_channels[flu_idx]]], na.rm = TRUE)
plt_flu <- ggplot2::ggplot(data = flu_norm_pr_data,
ggplot2::aes(x = .data$column, y = row, fill = .data[[flu_channels[flu_idx]]])) +
ggplot2::geom_tile() +
ggplot2::scale_x_discrete("", position = "top", limits = factor(unique(flu_norm_pr_data$column))) +
ggplot2::scale_y_discrete("", limits = rev(unique(flu_norm_pr_data$row))) +
viridis::scale_fill_viridis(paste0(flu_labels[flu_idx], " (rfu)"),
discrete = FALSE, limits = c(0, max_value),
alpha = 0.4,
na.value = "white") +
ggplot2::geom_text(ggplot2::aes(label = round(.data[[flu_channels[flu_idx]]], 2)), na.rm = TRUE, size = 2.5) +
ggplot2::theme_bw() + # base_size = 8
ggplot2::theme(
aspect.ratio = 8/12,
panel.grid = ggplot2::element_blank()
)
plt_flu
plotname <- paste0(flu_labels[flu_idx], "_1a_raw.pdf")
if(isFALSE(csv_only)){
ggplot2::ggsave(file.path(outfolder, plotname),
plot = plt_flu,
height = 16, width = 24, units = "cm")
}
# heatmap2 - normalised fluor
max_value <- max(flu_norm_pr_data[[paste0("normalised_", flu_channels[flu_idx])]], na.rm = TRUE)
plt_flu <- ggplot2::ggplot(data = flu_norm_pr_data,
ggplot2::aes(x = .data$column, y = row, fill = .data[[paste("normalised_", flu_channels[flu_idx], sep = "")]])) +
ggplot2::geom_tile() +
ggplot2::scale_x_discrete("", position = "top", # put axis labels at top
limits = factor(unique(flu_norm_pr_data$column))) + # put first element at top # factor to make it discrete rather than continuous
ggplot2::scale_y_discrete("", limits = rev(unique(flu_norm_pr_data$row))) + # put first element at top
viridis::scale_fill_viridis(paste0("normalised ", flu_labels[flu_idx], " (rfu)"),
discrete = FALSE, limits = c(0, max_value),
alpha = 0.4,
na.value = "white") +
ggplot2::geom_text(ggplot2::aes(label = round(.data[[paste("normalised_", flu_channels[flu_idx], sep = "")]], 2)),
na.rm = TRUE, size = 2.5) +
# ggplot2::coord_fixed(ratio = 1) +
ggplot2::theme_bw() + # base_size = 8
ggplot2::theme(
aspect.ratio = 8/12,
panel.grid = ggplot2::element_blank()
)
plt_flu
plotname <- paste0(flu_labels[flu_idx], "_1b_normalised.pdf")
if(isFALSE(csv_only)){
ggplot2::ggsave(file.path(outfolder, plotname),
plot = plt_flu,
height = 16, width = 24, units = "cm")
}
} else if(isTRUE(timecourse)){
plt_flu <- ggplot2::ggplot(flu_norm_pr_data) +
ggplot2::geom_line(ggplot2::aes(x = .data$time, y = .data[[flu_channels[flu_idx]]]),
colour = "black", linewidth = 0.5) +
ggplot2::geom_line(ggplot2::aes(x = .data$time,
y = .data[[paste("normalised_", flu_channels[flu_idx], sep = "")]]),
colour = "red", linewidth = 0.5) +
ggplot2::scale_x_continuous("time") +
ggplot2::scale_y_continuous(name = paste0(flu_labels[flu_idx], " (rfu)")) + # FP name
ggplot2::labs(caption = "black: raw, red: normalised") +
ggplot2::scale_colour_discrete("") +
ggplot2::facet_grid(row~column) +
ggplot2::theme_bw(base_size = 8) +
ggplot2::theme(
aspect.ratio = 1,
legend.position = "none",
axis.text.x = ggplot2::element_text(angle = 90, vjust = 0.5),
panel.grid.minor = ggplot2::element_blank()
)
plt_flu
plotname <- paste0(flu_labels[flu_idx], "_1_raw_normalised.pdf")
if(isFALSE(csv_only)){
ggplot2::ggsave(file.path(outfolder, plotname),
plot = plt_flu,
height = 16, width = 24, units = "cm")
}
}
}
out_data <- flu_norm_pr_data # copy normalised data to out_data df that will be returned if do_calibrate = FALSE
# Correct for cell quenching ----------------------------------------------------------------------
# Force od_name = NULL data to be do_quench_correction = FALSE (if by accident it isn't)
if(is.null(od_name) & isTRUE(do_quench_correction)){
message("Quench correction cannot be run without OD data. Setting do_quench_correction to FALSE and skipping this step...")
do_quench_correction <- FALSE
}
if (do_quench_correction) {
### Correct for cell quenching
## for each fluorescence channel...
for (flu_idx in seq_len(length(flu_channels))) {
out_data <- fpcountr::correct_flu(pr_data = out_data,
od_type = od_type,
flu_channel = flu_channels[flu_idx])
# plot calibrated fluorescence data
if(isFALSE(timecourse)){
# heatmap - corrected
max_value <- max(out_data[[paste0("corrected_normalised_", flu_channels[flu_idx])]], na.rm = TRUE)
plt_flu <- ggplot2::ggplot(data = out_data,
ggplot2::aes(x = .data$column, y = row, fill = .data[[paste("corrected_normalised_", flu_channels[flu_idx], sep = "")]])) +
ggplot2::geom_tile() +
ggplot2::scale_x_discrete("", position = "top", limits = factor(unique(out_data$column))) +
ggplot2::scale_y_discrete("", limits = rev(unique(out_data$row))) +
viridis::scale_fill_viridis(paste0("corrected ", flu_labels[flu_idx], " (rfu)"),
discrete = FALSE, limits = c(0, max_value),
alpha = 0.4,
na.value = "white") +
ggplot2::geom_text(ggplot2::aes(label = round(.data[[paste("corrected_normalised_", flu_channels[flu_idx], sep = "")]], 2)),
na.rm = TRUE, size = 2.5) +
ggplot2::theme_bw() + # base_size = 8
ggplot2::theme(
aspect.ratio = 8/12,
panel.grid = ggplot2::element_blank()
)
plt_flu
} else if(isTRUE(timecourse)){
plt_flu <- ggplot2::ggplot(out_data) +
ggplot2::geom_line(ggplot2::aes(x = .data$time,
y = .data[[paste("normalised_", flu_channels[flu_idx], sep = "")]],
colour = "normalised"),
colour = "black", linewidth = 0.5) +
ggplot2::geom_line(ggplot2::aes(x = .data$time,
y = .data[[paste("corrected_normalised_", flu_channels[flu_idx], sep = "")]],
colour = "corrected"),
colour = "red", linewidth = 0.5) +
ggplot2::scale_x_continuous("time") +
ggplot2::scale_y_continuous(name = paste0("corrected ", flu_labels[flu_idx], " (rfu)")) + # FP name
ggplot2::labs(caption = "black: normalised, red: corrected") +
ggplot2::scale_colour_discrete("") +
ggplot2::facet_grid(row~column) +
ggplot2::theme_bw(base_size = 8) +
ggplot2::theme(
aspect.ratio = 1,
legend.position = "none",
axis.text.x = ggplot2::element_text(angle = 90, vjust = 0.5),
panel.grid.minor = ggplot2::element_blank()
)
plt_flu
}
plotname <- paste0(flu_labels[flu_idx], "_2_quench-corrected.pdf")
if(isFALSE(csv_only)){
ggplot2::ggsave(file.path(outfolder, plotname),
plot = plt_flu,
height = 16, width = 24, units = "cm")
}
} # for each fluorescence channel
} # if do correct
# Calibrate data ----------------------------------------------------------------------
if (do_calibrate) {
if(is.null(od_name)){
# skip OD calibration...
# rename data done at calibrate_od() step - n/a
# NB. data here won't have calibrated_OD column
} else {
# proceed w OD calibration...
### Calibrate OD to Particle Number ----------------------------------------------------
out_data <- fpcountr::calibrate_od(pr_data = out_data,
od_name = od_name,
instr = instr,
conversion_factors_csv = od_coeffs_csv)
utils::head(out_data)
# adds calibrated_OD column
# plot
if(isFALSE(timecourse)){
# heatmap - calibrated OD
max_value <- max(out_data$calibrated_OD, na.rm = TRUE)
plt_od_calib <- ggplot2::ggplot(data = out_data,
ggplot2::aes(x = .data$column, y = row, fill = .data$calibrated_OD)) +
ggplot2::geom_tile() +
ggplot2::scale_x_discrete("", position = "top", limits = factor(unique(out_data$column))) +
ggplot2::scale_y_discrete("", limits = rev(unique(out_data$row))) +
viridis::scale_fill_viridis(paste0("calibrated ", od_name, " (particles)"),
discrete = FALSE, limits = c(0, max_value),
alpha = 0.4,
na.value = "white") +
ggplot2::geom_text(ggplot2::aes(label = scales::scientific(.data$calibrated_OD)), na.rm = TRUE, size = 2.5) +
ggplot2::theme_bw() + # base_size = 8
ggplot2::theme(
aspect.ratio = 8/12,
panel.grid = ggplot2::element_blank()
)
plt_od_calib
} else if(isTRUE(timecourse)){
plt_od_calib <- ggplot2::ggplot(out_data) +
ggplot2::geom_line(ggplot2::aes(x = .data$time, y = .data$calibrated_OD,
colour = "calibrated"), linewidth = 0.5) +
ggplot2::scale_x_continuous("time") +
ggplot2::scale_y_continuous(name = paste0("calibrated ", od_name, " (particles)")) + # OD700 (particles)
ggplot2::labs(caption = "") +
ggplot2::scale_colour_discrete("") +
ggplot2::facet_grid(row~column) +
ggplot2::theme_bw(base_size = 8) +
ggplot2::theme(
aspect.ratio = 1,
legend.position = "none",
axis.text.x = ggplot2::element_text(angle = 90, vjust = 0.5),
panel.grid.minor = ggplot2::element_blank()
)
plt_od_calib
}
plotname <- "OD_3_calibrated.pdf"
if(isFALSE(csv_only)){
ggplot2::ggsave(file.path(outfolder, plotname),
plot = plt_od_calib,
height = 16, width = 24, units = "cm")
}
} # od calibration
### Calibrate RFU to Molecule number ----------------------------------------------------
## for each fluorescence channel...
for (flu_idx in seq_len(length(flu_channels))) {
# original: 'as long as the current index is within the few that you've decided to calibrate....'
# removing this bc here, all fluorescence channels ought to be calibratable!
# if(length(flu_gains) >= flu_idx){
out_data <- fpcountr::calibrate_flu(pr_data = out_data,
flu_instr = instr,
flu_channel = flu_channels[flu_idx],
flu_gain = flu_gains[flu_idx],
flu_slug = flu_slugs[flu_idx],
flu_label = flu_labels[flu_idx],
do_quench_correction = do_quench_correction,
conversion_factors_csv = fluor_coeffs_csv)
# plot calibrated fluorescence data
if(isFALSE(timecourse)){
# heatmap - calibrated OD
max_value <- max(out_data[[paste0("calibrated_", flu_labels[flu_idx])]], na.rm = TRUE)
plt_flu_calib <- ggplot2::ggplot(data = out_data,
ggplot2::aes(x = .data$column, y = row, fill = .data[[paste("calibrated_", flu_labels[flu_idx], sep = "")]])) +
ggplot2::geom_tile() +
ggplot2::scale_x_discrete("", position = "top", limits = factor(unique(out_data$column))) +
ggplot2::scale_y_discrete("", limits = rev(unique(out_data$row))) +
viridis::scale_fill_viridis(paste0(flu_labels[flu_idx], " (molecules)"),
discrete = FALSE, limits = c(0, max_value),
alpha = 0.4,
na.value = "white") +
ggplot2::geom_text(ggplot2::aes(label = scales::scientific(.data[[paste("calibrated_", flu_labels[flu_idx], sep = "")]])), na.rm = TRUE, size = 2.5) +
ggplot2::theme_bw() + # base_size = 8
ggplot2::theme(
aspect.ratio = 8/12,
panel.grid = ggplot2::element_blank()
)
plt_flu_calib
} else if(isTRUE(timecourse)){
plt_flu_calib <- ggplot2::ggplot(out_data) +
ggplot2::geom_line(ggplot2::aes(x = .data$time,
y = .data[[paste("calibrated_", flu_labels[flu_idx], sep = "")]],
colour = "calibrated"),
linewidth = 0.5) +
ggplot2::scale_x_continuous("time") +
ggplot2::scale_y_continuous(name = paste0(flu_labels[flu_idx], " (molecules)")) + # FP name
ggplot2::labs(caption = "") +
ggplot2::scale_colour_discrete("") +
ggplot2::facet_grid(row~column) +
ggplot2::theme_bw(base_size = 8) +
ggplot2::theme(
aspect.ratio = 1,
legend.position = "none",
axis.text.x = ggplot2::element_text(angle = 90, vjust = 0.5),
panel.grid.minor = ggplot2::element_blank()
)
plt_flu_calib
}
plotname <- paste0(flu_labels[flu_idx], "_3_calibrated.pdf")
if(isFALSE(csv_only)){
ggplot2::ggsave(file.path(outfolder, plotname),
plot = plt_flu_calib,
height = 16, width = 24, units = "cm")
}
# original: 'as long as the current index is within the few that you've decided to calibrate....'
# removing this bc here, all fluorescence channels ought to be calibratable!
# } # if(length(flu_gains) >= flu_idx){
# else {break} # acts to break if loop once number exceeds number of flu_channels
# } # if else
} # for each fluorescence channel
} # if do calibrate
# Save ------------------------------------------------------------------------------------------------
filename <- gsub(".csv", "_processed.csv", basename(data_csv))
utils::write.csv(x = out_data, file = file.path(outfolder, filename), row.names = FALSE)
return(out_data)
}
ec363/fpcountr documentation built on Nov. 29, 2024, 12:03 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 process_plate
Documented in process_plate
#' Process experimental data (normalise, correct and calibrate)
#'
#' Processes experimental (fluorescence and absorbance) data from microplate
#' readers. Expects 'parsed' data that is tidy and attached to appropriate
#' metadata. Numerous arguments offer options to customise how and whether
#' normalisation, correction and/or calibration is carried out. Data can be
#' timecourse (kinetic) data or single timepoint (endpoint) data, and can
#' contain fluorescence data only or include absorbance (OD) data to enumerate
#' cell density. Originally based on `flopr::process_plate()` but with numerous
#' changes. First, the function normalises optical density readings (if present)
#' to media/buffer wells, and fluorescence readings, either to media/buffer
#' wells or to cellular autofluorescence. Fluorescence channels must be
#' specified in `flu_channels`, the blank wells in `blank_wells`, and the
#' reading to be used for optical density in `od_name`. The cellular
#' autofluorescence model can be set with `af_model` on the negative
#' (`neg_well`) wells, or set to `NULL`. Second, if `do_quench_correction` is
#' TRUE, the 'quenching' or attenuation of fluorescence measurements by cells is
#' corrected for according to the cell density measured. This requires the
#' specification of `od_type` as "OD600" or "OD700". Third, if `do_calibrate` is
#' TRUE, normalised values for both fluorescence and optical density are
#' converted into meaningful units. Here, the paths to conversion factor CSV
#' files must be provided as `od_coeffs_csv` for OD and `fluor_coeffs_csv` for
#' fluorescence. For fluorescence measurements, each calibration must be
#' specified by `flu_slugs` to represent the FP used, `flu_gains` to provide the
#' gain, and `flu_labels` to specify how the relevant plots should be labelled,
#' (e.g. `flu_slugs = c("mcherry", "mtagbfp2")`, `flu_gains = c(60,80)`,
#' `flu_labels = c("RFP, BFP")`).
#'
#' @param data_csv path to a CSV file containing parsed plate reader data
#' @param blank_well the well coordinates of one or more media blanks. Defaults
#' to "A1".
#' @param timecourse logical. Is the data timecourse/kinetic data and does it
#' include a variable called 'time'?
#' @param od_name the column name for the optical density data. Defaults to
#' "OD". If no OD measurements were taken, use `NULL`.
#' @param flu_channels the column names for the fluorescence data. Defaults to
#' "green1green2".
#' @param flu_channels_rename if specified, what to change the flu_channels
#' column names to, before computing normalisations and calibrations. Can be
#' useful if the channel_name in the calibration file is named differently
#' from the columns in the data file. Needs to be same length as flu_channels,
#' if not all require changing, specify them anyway to allow positional
#' replacement (first element in flu_channels_rename replaces first in
#' flu_channels, etc.). Defaults to NULL.
#' @param af_model model used to fit negative control autofluorescence. For now
#' these include "polynomial", "inverse_poly", "exponential", "spline" and
#' "loess". If set to NULL, no model is used, and fluorescence is normalised
#' akin to OD: by subtracting the value for the blanks. Defaults to "spline".
#' @param neg_well the well coordinates of a non-fluorescent control. Defaults
#' to "A2".
#' @param do_quench_correction logical. Should function correct for anticipated
#' quenching of fluorescence, depending on the cell density?
#' @param od_type Which OD-type was used? Required for quench correction.
#' "OD600" or "OD700".
#' @param do_calibrate logical. Should function convert OD and fluorescence data
#' to calibrated units? Defaults to FALSE.
#' @param instr character string to represent instrument. If do_calibrate =
#' TRUE, used for filtering `od_coeffs_csv` and `fluor_coeffs_csv` files for
#' conversion factors of the relevant instrument.
#' @param flu_slugs character array representing fluorescent proteins (format =
#' FPbase slug). If do_calibrate = TRUE, used for filtering `fluor_coeffs_csv`
#' for conversion factors of the relevant FP.
#' @param flu_gains numeric array representing gains of each fluorescent
#' channel. If do_calibrate = TRUE, used for filtering `fluor_coeffs_csv` for
#' conversion factors of the relevant gain.
#' @param flu_labels If do_calibrate = TRUE, the column names to be given to
#' each calibration. May be identical to flu_slug or flu_channel, but
#' recommended is to make it obvious which FP is being calibrated, e.g.
#' "mCherry", as channel names may be non-specifically named e.g. "red1red1".
#' Needs to be same length as flu_slugs and flu_gains.
#' @param od_coeffs_csv if do_calibrate = TRUE, path of the CSV file containing
#' conversion factors for optical density
#' @param fluor_coeffs_csv if do_calibrate = TRUE, path of the CSV file
#' containing predicted conversion factors for the fluorescent channels
#' @param outfolder path to folder where output files should be saved. Defaults
#' to current working directory.
#' @param csv_only logical. Saves only CSV files as outputs when `TRUE`.
#' Defaults to `FALSE`.
#'
#' @return a data.frame with columns for raw plate reader data, normalised data
#' and, if do_calibrate = TRUE, calibrated OD and FP data
#' @export
#' @importFrom rlang .data
#'
#' @examples
#' \dontrun{
#' processed_data <- process_plate(
#' data_csv = "mcherry_parsed.csv",
#' blank_well = c("A11"),
#' od_name = "OD600", flu_channels = c("red1"), flu_channels_rename = c("red1red1"),
#' af_model = NULL,
#' do_quench_correction = TRUE, od_type = "OD700",
#' do_calibrate = TRUE, instr = "spark1",
#' flu_slugs = c("mcherry"), flu_gains = c(80), flu_labels = c("mcherry"),
#' od_coeffs_csv = "od_coeffs.csv", fluor_coeffs_csv = "flu_coeffs.csv",
#' outfolder = file.path("data_processed")
#' )
#' }
process_plate <- function(
data_csv, blank_well = "A1",
# timecourse
timecourse = TRUE,
# od
od_name = "OD",
# fluorescence
flu_channels = c("green1green2"), # column names in data
flu_channels_rename = NULL, # if not NULL, for every flu_channels specify new name
# rename data columns to make sure they match conversion factor table entries
# autofluorescence model
af_model = "spline", # options: NULL, "spline", "loess"
neg_well = "A2", # if af_model = NULL this can be hashed
# cell quench correction
do_quench_correction = FALSE,
od_type, # "OD600" or "OD700"
# calibrations
do_calibrate = FALSE,
instr, # instrument name for the calibration(s)
# flu_channels above specifies channel(s) for the calibration(s)
flu_slugs = c(), # FP name in FPbase's slug format for the calibration(s)
# OR whatever matches the format used in the conversion factor tables
flu_gains = c(), # gain for calibration(s)
# each of the above need to be specified for every flu_channels being calibrated
flu_labels = c(), # how to display the FP/channel combination in plots and file names
# important if slug is unhelpful or multiple gains used for same FP
# conversion factors
od_coeffs_csv, # microsphere conversion factors for od600, od700
fluor_coeffs_csv, # FP conversion factors
outfolder = ".",
csv_only = FALSE
) {
# Get parsed data --------------------------------------------------------------------------------------------------
pr_data <- utils::read.csv(data_csv, check.names = FALSE)
pr_data
# Rename flu channels if requested ---------------------------------------------------------------------------------
if(!is.null(flu_channels_rename)){
## Check there are replacements for each flu channel
if(!isTRUE(all.equal(length(flu_channels), length(flu_channels_rename), check.attributes = FALSE))){
message("Error: number of flu_channels must match number of flu_channels_rename.
To remove requirement for renaming, remove flu_channels_rename or set it to NULL.")
return()
}
## Rename columns
for (flu_name_idx in seq_len(length(flu_channels))){
names(pr_data)
col_idx <- which(names(pr_data) == flu_channels[flu_name_idx])
names(pr_data)[col_idx] <- flu_channels_rename[flu_name_idx]
names(pr_data)
}
# Rest of script thinks flu_channels are the final column names needed.
# So flu_channels needs to be overwritten with flu_channels_rename to prevent downstream errors
flu_channels <- flu_channels_rename
}
# Location for saved plots -----------------------------------------------------------------------------------------
# make folder if it doesn't exist already
ifelse(test = !dir.exists(file.path(outfolder)), yes = dir.create(file.path(outfolder)), no = FALSE)
# OD processing sections --------------------------------------------------------------------
if(is.null(od_name)){
# skip OD processing...
# rename data
od_norm_pr_data <- pr_data # otherwise done at od_norm() step # NB. this won't have normalised_OD or normalised_OD_cm1 columns
} else {
# proceed with OD processing...
# Normalise OD (with external od_norm function) --------------------------------------------------------------------
od_norm_pr_data <- fpcountr::od_norm(pr_data, blank_well, od_name, timecourse = timecourse)
# adds one extra column - normalised_OD. for each timepoint, normalises OD to mean of blank wells
# Plot OD data in plate format
# plot
if(isFALSE(timecourse)){
# heatmap1 - raw OD
max_value <- max(od_norm_pr_data[[od_name]], na.rm = TRUE)
plt_od <- ggplot2::ggplot(data = od_norm_pr_data,
# ggplot2::aes(x = row, y = column, fill = .data[[od_name]])) +
ggplot2::aes(x = .data$column, y = row, fill = .data[[od_name]])) +
ggplot2::geom_tile() +
ggplot2::scale_x_discrete("", position = "top", # put axis labels at top
limits = factor(unique(od_norm_pr_data$column))) + # put first element at top # factor to make it discrete rather than continuous
ggplot2::scale_y_discrete("", limits = rev(unique(od_norm_pr_data$row))) + # put first element at top
viridis::scale_fill_viridis("raw OD",
discrete = FALSE, limits = c(0, max_value),
alpha = 0.4,
na.value = "white") +
ggplot2::geom_text(ggplot2::aes(label = round(.data[[od_name]], 2)), na.rm = TRUE, size = 2.5) +
# ggplot2::coord_fixed(ratio = 1) +
ggplot2::theme_bw() + # base_size = 8
ggplot2::theme(
aspect.ratio = 8/12,
panel.grid = ggplot2::element_blank()
)
plt_od
plotname <- "OD_1a_raw.pdf"
if(isFALSE(csv_only)){
ggplot2::ggsave(file.path(outfolder, plotname),
plot = plt_od,
height = 16, width = 24, units = "cm")
}
# normalised OD
max_value <- max(od_norm_pr_data$normalised_OD, na.rm = TRUE)
plt_od <- ggplot2::ggplot(data = od_norm_pr_data,
ggplot2::aes(x = .data$column, y = row, fill = .data$normalised_OD)) +
ggplot2::geom_tile() +
ggplot2::scale_x_discrete("", position = "top", limits = factor(unique(od_norm_pr_data$column))) +
ggplot2::scale_y_discrete("", limits = rev(unique(od_norm_pr_data$row))) +
viridis::scale_fill_viridis("normalised OD",
discrete = FALSE, limits = c(0, max_value),
alpha = 0.4,
na.value = "white") +
ggplot2::geom_text(ggplot2::aes(label = round(.data$normalised_OD, 2)), na.rm = TRUE, size = 2.5) +
ggplot2::theme_bw() + # base_size = 8
ggplot2::theme(
aspect.ratio = 8/12,
panel.grid = ggplot2::element_blank()
)
plt_od
plotname <- "OD_1b_normalised.pdf"
if(isFALSE(csv_only)){
ggplot2::ggsave(file.path(outfolder, plotname),
plot = plt_od,
height = 16, width = 24, units = "cm")
}
} else if(isTRUE(timecourse)){
# plot with caption
plt_od <- ggplot2::ggplot(od_norm_pr_data) +
ggplot2::geom_line(ggplot2::aes(x = .data$time, y = .data[[od_name]]),
colour = "black", linewidth = 0.5) +
ggplot2::geom_line(ggplot2::aes(x = .data$time, y = .data$normalised_OD),
colour = "red", linewidth = 0.5) +
ggplot2::scale_x_continuous("time") +
ggplot2::labs(caption = "black: raw, red: normalised") +
ggplot2::scale_colour_discrete("") +
ggplot2::facet_grid(row~column) +
ggplot2::theme_bw(base_size = 8) +
ggplot2::theme(
aspect.ratio = 1,
legend.position = "none",
axis.text.x = ggplot2::element_text(angle = 90, vjust = 0.5),
panel.grid.minor = ggplot2::element_blank()
)
plt_od
plotname <- "OD_1_raw_normalised.pdf"
if(isFALSE(csv_only)){
ggplot2::ggsave(file.path(outfolder, plotname),
plot = plt_od,
height = 16, width = 24, units = "cm")
}
}
# Convert to OD (cm-1) --------------------------------------------------------------------
if("volume" %in% names(od_norm_pr_data)){ # only run if data contains the column 'volume'
od_norm_pr_data <- od_norm_pr_data %>%
dplyr::mutate(pathlength = fpcountr::get_pathlength(.data$volume)) %>% # get (estimated) pathlength from volume
dplyr::mutate(normalised_OD_cm1 = .data$normalised_OD/.data$pathlength) # get OD (cm-1)
od_norm_pr_data[1,]
# Plot in plate format
if(isFALSE(timecourse)){
# heatmap1 - raw OD
max_value <- max(od_norm_pr_data$normalised_OD_cm1, na.rm = TRUE)
plt_od <- ggplot2::ggplot(data = od_norm_pr_data,
ggplot2::aes(x = .data$column, y = row, fill = .data$normalised_OD_cm1)) +
ggplot2::geom_tile() +
ggplot2::scale_x_discrete("", position = "top", limits = factor(unique(od_norm_pr_data$column))) +
ggplot2::scale_y_discrete("", limits = rev(unique(od_norm_pr_data$row))) +
viridis::scale_fill_viridis(paste0("normalised ", od_name, " (cm-1)"),
discrete = FALSE, limits = c(0, max_value),
alpha = 0.4,
na.value = "white") +
ggplot2::geom_text(ggplot2::aes(label = round(.data$normalised_OD_cm1, 2)), na.rm = TRUE, size = 2.5) +
ggplot2::theme_bw() + # base_size = 8
ggplot2::theme(
aspect.ratio = 8/12,
panel.grid = ggplot2::element_blank()
)
plt_od
} else if(isTRUE(timecourse)){
plt_od <- ggplot2::ggplot(od_norm_pr_data) +
ggplot2::geom_line(ggplot2::aes(x = .data$time, y = .data$normalised_OD_cm1,
colour = "normalised_cm1"), linewidth = 0.5) +
ggplot2::scale_x_continuous("time") +
ggplot2::scale_y_continuous(name = paste0("normalised ", od_name, " (cm-1)")) + # "normalised OD (cm-1)"
ggplot2::labs(caption = "") +
ggplot2::scale_colour_discrete("") +
ggplot2::facet_grid(row~column) +
ggplot2::theme_bw(base_size = 8) +
ggplot2::theme(
aspect.ratio = 1,
legend.position = "none",
axis.text.x = ggplot2::element_text(angle = 90, vjust = 0.5),
panel.grid.minor = ggplot2::element_blank()
)
plt_od
}
plotname <- "OD_2_pathlength-normalised.pdf"
if(isFALSE(csv_only)){
ggplot2::ggsave(file.path(outfolder, plotname),
plot = plt_od,
height = 16, width = 24, units = "cm")
}
}
} # od processing sections
# Normalise fluorescence data --------------------------------------------------------------------------------------------
flu_norm_pr_data <- od_norm_pr_data
## for each fluorescence channel...
for (flu_idx in seq_len(length(flu_channels))) {
## for each fluorescent protein, run fluorescent normalisation function..
# Force od_name = NULL data to be af_model = NULL (if by accident it isn't)
if(is.null(od_name) & !is.null(af_model)){
message("Autofluorescence model function cannot be run without OD data. Setting af_model to NULL and normalising fluorescence to blank_wells...")
af_model <- NULL
}
# Force timecourse = FALSE data to be af_model = NULL (if by accident it isn't)
if(isFALSE(timecourse) & !is.null(af_model)){
message("Autofluorescence model function cannot be run without timecourse data. Setting af_model to NULL and normalising fluorescence to blank_wells...")
af_model <- NULL
}
flu_norm_pr_data <- fpcountr::flu_norm(pr_data = flu_norm_pr_data, neg_well = neg_well, blank_well = blank_well,
flu_name = flu_channels[flu_idx], af_model = af_model, data_csv = data_csv,
timecourse = timecourse, outfolder = outfolder)
# adds one column per FP of (eg) normalised_[GFP] to the table
# Plot in plate format
if(isFALSE(timecourse)){
# heatmap1 - raw fluor
max_value <- max(flu_norm_pr_data[[flu_channels[flu_idx]]], na.rm = TRUE)
plt_flu <- ggplot2::ggplot(data = flu_norm_pr_data,
ggplot2::aes(x = .data$column, y = row, fill = .data[[flu_channels[flu_idx]]])) +
ggplot2::geom_tile() +
ggplot2::scale_x_discrete("", position = "top", limits = factor(unique(flu_norm_pr_data$column))) +
ggplot2::scale_y_discrete("", limits = rev(unique(flu_norm_pr_data$row))) +
viridis::scale_fill_viridis(paste0(flu_labels[flu_idx], " (rfu)"),
discrete = FALSE, limits = c(0, max_value),
alpha = 0.4,
na.value = "white") +
ggplot2::geom_text(ggplot2::aes(label = round(.data[[flu_channels[flu_idx]]], 2)), na.rm = TRUE, size = 2.5) +
ggplot2::theme_bw() + # base_size = 8
ggplot2::theme(
aspect.ratio = 8/12,
panel.grid = ggplot2::element_blank()
)
plt_flu
plotname <- paste0(flu_labels[flu_idx], "_1a_raw.pdf")
if(isFALSE(csv_only)){
ggplot2::ggsave(file.path(outfolder, plotname),
plot = plt_flu,
height = 16, width = 24, units = "cm")
}
# heatmap2 - normalised fluor
max_value <- max(flu_norm_pr_data[[paste0("normalised_", flu_channels[flu_idx])]], na.rm = TRUE)
plt_flu <- ggplot2::ggplot(data = flu_norm_pr_data,
ggplot2::aes(x = .data$column, y = row, fill = .data[[paste("normalised_", flu_channels[flu_idx], sep = "")]])) +
ggplot2::geom_tile() +
ggplot2::scale_x_discrete("", position = "top", # put axis labels at top
limits = factor(unique(flu_norm_pr_data$column))) + # put first element at top # factor to make it discrete rather than continuous
ggplot2::scale_y_discrete("", limits = rev(unique(flu_norm_pr_data$row))) + # put first element at top
viridis::scale_fill_viridis(paste0("normalised ", flu_labels[flu_idx], " (rfu)"),
discrete = FALSE, limits = c(0, max_value),
alpha = 0.4,
na.value = "white") +
ggplot2::geom_text(ggplot2::aes(label = round(.data[[paste("normalised_", flu_channels[flu_idx], sep = "")]], 2)),
na.rm = TRUE, size = 2.5) +
# ggplot2::coord_fixed(ratio = 1) +
ggplot2::theme_bw() + # base_size = 8
ggplot2::theme(
aspect.ratio = 8/12,
panel.grid = ggplot2::element_blank()
)
plt_flu
plotname <- paste0(flu_labels[flu_idx], "_1b_normalised.pdf")
if(isFALSE(csv_only)){
ggplot2::ggsave(file.path(outfolder, plotname),
plot = plt_flu,
height = 16, width = 24, units = "cm")
}
} else if(isTRUE(timecourse)){
plt_flu <- ggplot2::ggplot(flu_norm_pr_data) +
ggplot2::geom_line(ggplot2::aes(x = .data$time, y = .data[[flu_channels[flu_idx]]]),
colour = "black", linewidth = 0.5) +
ggplot2::geom_line(ggplot2::aes(x = .data$time,
y = .data[[paste("normalised_", flu_channels[flu_idx], sep = "")]]),
colour = "red", linewidth = 0.5) +
ggplot2::scale_x_continuous("time") +
ggplot2::scale_y_continuous(name = paste0(flu_labels[flu_idx], " (rfu)")) + # FP name
ggplot2::labs(caption = "black: raw, red: normalised") +
ggplot2::scale_colour_discrete("") +
ggplot2::facet_grid(row~column) +
ggplot2::theme_bw(base_size = 8) +
ggplot2::theme(
aspect.ratio = 1,
legend.position = "none",
axis.text.x = ggplot2::element_text(angle = 90, vjust = 0.5),
panel.grid.minor = ggplot2::element_blank()
)
plt_flu
plotname <- paste0(flu_labels[flu_idx], "_1_raw_normalised.pdf")
if(isFALSE(csv_only)){
ggplot2::ggsave(file.path(outfolder, plotname),
plot = plt_flu,
height = 16, width = 24, units = "cm")
}
}
}
out_data <- flu_norm_pr_data # copy normalised data to out_data df that will be returned if do_calibrate = FALSE
# Correct for cell quenching ----------------------------------------------------------------------
# Force od_name = NULL data to be do_quench_correction = FALSE (if by accident it isn't)
if(is.null(od_name) & isTRUE(do_quench_correction)){
message("Quench correction cannot be run without OD data. Setting do_quench_correction to FALSE and skipping this step...")
do_quench_correction <- FALSE
}
if (do_quench_correction) {
### Correct for cell quenching
## for each fluorescence channel...
for (flu_idx in seq_len(length(flu_channels))) {
out_data <- fpcountr::correct_flu(pr_data = out_data,
od_type = od_type,
flu_channel = flu_channels[flu_idx])
# plot calibrated fluorescence data
if(isFALSE(timecourse)){
# heatmap - corrected
max_value <- max(out_data[[paste0("corrected_normalised_", flu_channels[flu_idx])]], na.rm = TRUE)
plt_flu <- ggplot2::ggplot(data = out_data,
ggplot2::aes(x = .data$column, y = row, fill = .data[[paste("corrected_normalised_", flu_channels[flu_idx], sep = "")]])) +
ggplot2::geom_tile() +
ggplot2::scale_x_discrete("", position = "top", limits = factor(unique(out_data$column))) +
ggplot2::scale_y_discrete("", limits = rev(unique(out_data$row))) +
viridis::scale_fill_viridis(paste0("corrected ", flu_labels[flu_idx], " (rfu)"),
discrete = FALSE, limits = c(0, max_value),
alpha = 0.4,
na.value = "white") +
ggplot2::geom_text(ggplot2::aes(label = round(.data[[paste("corrected_normalised_", flu_channels[flu_idx], sep = "")]], 2)),
na.rm = TRUE, size = 2.5) +
ggplot2::theme_bw() + # base_size = 8
ggplot2::theme(
aspect.ratio = 8/12,
panel.grid = ggplot2::element_blank()
)
plt_flu
} else if(isTRUE(timecourse)){
plt_flu <- ggplot2::ggplot(out_data) +
ggplot2::geom_line(ggplot2::aes(x = .data$time,
y = .data[[paste("normalised_", flu_channels[flu_idx], sep = "")]],
colour = "normalised"),
colour = "black", linewidth = 0.5) +
ggplot2::geom_line(ggplot2::aes(x = .data$time,
y = .data[[paste("corrected_normalised_", flu_channels[flu_idx], sep = "")]],
colour = "corrected"),
colour = "red", linewidth = 0.5) +
ggplot2::scale_x_continuous("time") +
ggplot2::scale_y_continuous(name = paste0("corrected ", flu_labels[flu_idx], " (rfu)")) + # FP name
ggplot2::labs(caption = "black: normalised, red: corrected") +
ggplot2::scale_colour_discrete("") +
ggplot2::facet_grid(row~column) +
ggplot2::theme_bw(base_size = 8) +
ggplot2::theme(
aspect.ratio = 1,
legend.position = "none",
axis.text.x = ggplot2::element_text(angle = 90, vjust = 0.5),
panel.grid.minor = ggplot2::element_blank()
)
plt_flu
}
plotname <- paste0(flu_labels[flu_idx], "_2_quench-corrected.pdf")
if(isFALSE(csv_only)){
ggplot2::ggsave(file.path(outfolder, plotname),
plot = plt_flu,
height = 16, width = 24, units = "cm")
}
} # for each fluorescence channel
} # if do correct
# Calibrate data ----------------------------------------------------------------------
if (do_calibrate) {
if(is.null(od_name)){
# skip OD calibration...
# rename data done at calibrate_od() step - n/a
# NB. data here won't have calibrated_OD column
} else {
# proceed w OD calibration...
### Calibrate OD to Particle Number ----------------------------------------------------
out_data <- fpcountr::calibrate_od(pr_data = out_data,
od_name = od_name,
instr = instr,
conversion_factors_csv = od_coeffs_csv)
utils::head(out_data)
# adds calibrated_OD column
# plot
if(isFALSE(timecourse)){
# heatmap - calibrated OD
max_value <- max(out_data$calibrated_OD, na.rm = TRUE)
plt_od_calib <- ggplot2::ggplot(data = out_data,
ggplot2::aes(x = .data$column, y = row, fill = .data$calibrated_OD)) +
ggplot2::geom_tile() +
ggplot2::scale_x_discrete("", position = "top", limits = factor(unique(out_data$column))) +
ggplot2::scale_y_discrete("", limits = rev(unique(out_data$row))) +
viridis::scale_fill_viridis(paste0("calibrated ", od_name, " (particles)"),
discrete = FALSE, limits = c(0, max_value),
alpha = 0.4,
na.value = "white") +
ggplot2::geom_text(ggplot2::aes(label = scales::scientific(.data$calibrated_OD)), na.rm = TRUE, size = 2.5) +
ggplot2::theme_bw() + # base_size = 8
ggplot2::theme(
aspect.ratio = 8/12,
panel.grid = ggplot2::element_blank()
)
plt_od_calib
} else if(isTRUE(timecourse)){
plt_od_calib <- ggplot2::ggplot(out_data) +
ggplot2::geom_line(ggplot2::aes(x = .data$time, y = .data$calibrated_OD,
colour = "calibrated"), linewidth = 0.5) +
ggplot2::scale_x_continuous("time") +
ggplot2::scale_y_continuous(name = paste0("calibrated ", od_name, " (particles)")) + # OD700 (particles)
ggplot2::labs(caption = "") +
ggplot2::scale_colour_discrete("") +
ggplot2::facet_grid(row~column) +
ggplot2::theme_bw(base_size = 8) +
ggplot2::theme(
aspect.ratio = 1,
legend.position = "none",
axis.text.x = ggplot2::element_text(angle = 90, vjust = 0.5),
panel.grid.minor = ggplot2::element_blank()
)
plt_od_calib
}
plotname <- "OD_3_calibrated.pdf"
if(isFALSE(csv_only)){
ggplot2::ggsave(file.path(outfolder, plotname),
plot = plt_od_calib,
height = 16, width = 24, units = "cm")
}
} # od calibration
### Calibrate RFU to Molecule number ----------------------------------------------------
## for each fluorescence channel...
for (flu_idx in seq_len(length(flu_channels))) {
# original: 'as long as the current index is within the few that you've decided to calibrate....'
# removing this bc here, all fluorescence channels ought to be calibratable!
# if(length(flu_gains) >= flu_idx){
out_data <- fpcountr::calibrate_flu(pr_data = out_data,
flu_instr = instr,
flu_channel = flu_channels[flu_idx],
flu_gain = flu_gains[flu_idx],
flu_slug = flu_slugs[flu_idx],
flu_label = flu_labels[flu_idx],
do_quench_correction = do_quench_correction,
conversion_factors_csv = fluor_coeffs_csv)
# plot calibrated fluorescence data
if(isFALSE(timecourse)){
# heatmap - calibrated OD
max_value <- max(out_data[[paste0("calibrated_", flu_labels[flu_idx])]], na.rm = TRUE)
plt_flu_calib <- ggplot2::ggplot(data = out_data,
ggplot2::aes(x = .data$column, y = row, fill = .data[[paste("calibrated_", flu_labels[flu_idx], sep = "")]])) +
ggplot2::geom_tile() +
ggplot2::scale_x_discrete("", position = "top", limits = factor(unique(out_data$column))) +
ggplot2::scale_y_discrete("", limits = rev(unique(out_data$row))) +
viridis::scale_fill_viridis(paste0(flu_labels[flu_idx], " (molecules)"),
discrete = FALSE, limits = c(0, max_value),
alpha = 0.4,
na.value = "white") +
ggplot2::geom_text(ggplot2::aes(label = scales::scientific(.data[[paste("calibrated_", flu_labels[flu_idx], sep = "")]])), na.rm = TRUE, size = 2.5) +
ggplot2::theme_bw() + # base_size = 8
ggplot2::theme(
aspect.ratio = 8/12,
panel.grid = ggplot2::element_blank()
)
plt_flu_calib
} else if(isTRUE(timecourse)){
plt_flu_calib <- ggplot2::ggplot(out_data) +
ggplot2::geom_line(ggplot2::aes(x = .data$time,
y = .data[[paste("calibrated_", flu_labels[flu_idx], sep = "")]],
colour = "calibrated"),
linewidth = 0.5) +
ggplot2::scale_x_continuous("time") +
ggplot2::scale_y_continuous(name = paste0(flu_labels[flu_idx], " (molecules)")) + # FP name
ggplot2::labs(caption = "") +
ggplot2::scale_colour_discrete("") +
ggplot2::facet_grid(row~column) +
ggplot2::theme_bw(base_size = 8) +
ggplot2::theme(
aspect.ratio = 1,
legend.position = "none",
axis.text.x = ggplot2::element_text(angle = 90, vjust = 0.5),
panel.grid.minor = ggplot2::element_blank()
)
plt_flu_calib
}
plotname <- paste0(flu_labels[flu_idx], "_3_calibrated.pdf")
if(isFALSE(csv_only)){
ggplot2::ggsave(file.path(outfolder, plotname),
plot = plt_flu_calib,
height = 16, width = 24, units = "cm")
}
# original: 'as long as the current index is within the few that you've decided to calibrate....'
# removing this bc here, all fluorescence channels ought to be calibratable!
# } # if(length(flu_gains) >= flu_idx){
# else {break} # acts to break if loop once number exceeds number of flu_channels
# } # if else
} # for each fluorescence channel
} # if do calibrate
# Save ------------------------------------------------------------------------------------------------
filename <- gsub(".csv", "_processed.csv", basename(data_csv))
utils::write.csv(x = out_data, file = file.path(outfolder, filename), row.names = FALSE)
return(out_data)
}
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.