R/plotTIEC.R
In mcalgaro93/benchdamic: Benchmark of differential abundance methods on microbiome data
Defines functions
plotKS
plotLogP
plotQQ
plotFDR
plotFPR
Documented in
plotFDR
plotFPR
plotKS
plotLogP
plotQQ
#' @title plotFPR
#'
#' @export
#' @importFrom plyr ddply
#' @importFrom reshape2 melt
#' @importFrom tidytext reorder_within scale_x_reordered
#' @import ggplot2
#' @description
#' Draw the boxplots of the proportions of p-values lower than 0.01, 0.05, and
#' 0.1 thresholds for each method.
#'
#' @param df_FPR a \code{data.frame} produced by the \code{\link{createTIEC}}
#' function, containing the FPR values.
#' @param cols named vector of colors.
#'
#' @return A ggplot object.
#'
#' @examples
#' # Load some data
#' data(ps_stool_16S)
#'
#' # Generate the patterns for 10 mock comparison for an experiment
#' # (N = 1000 is suggested)
#' mocks <- createMocks(nsamples = phyloseq::nsamples(ps_stool_16S), N = 10)
#' head(mocks)
#'
#' # Add some normalization/scaling factors to the phyloseq object
#' my_norm <- setNormalizations(fun = c("norm_edgeR", "norm_CSS"),
#' method = c("TMM", "CSS"))
#' ps_stool_16S <- runNormalizations(normalization_list = my_norm,
#' object = ps_stool_16S)
#'
#' # Initialize some limma based methods
#' my_limma <- set_limma(design = ~ group, coef = 2,
#' norm = c("TMM", "CSS"))
#'
#' # Run methods on mock datasets
#' results <- runMocks(mocks = mocks, method_list = my_limma,
#' object = ps_stool_16S)
#'
#' # Prepare results for Type I Error Control
#' TIEC_summary <- createTIEC(results)
#'
#' # Plot the results
#' plotFPR(df_FPR = TIEC_summary$df_FPR)
#' plotFDR(df_FDR = TIEC_summary$df_FDR)
#' plotQQ(df_QQ = TIEC_summary$df_QQ, zoom = c(0, 0.1))
#' plotKS(df_KS = TIEC_summary$df_KS)
#' plotLogP(df_QQ = TIEC_summary$df_QQ)
plotFPR <- function(df_FPR, cols = NULL) {
# Create the vector of colors
if (is.null(cols)) {
cols <- createColors(variable = df_FPR$Method)
}
if (is.null(names(cols))) {
stop("'cols' vector is not a named vector of colors.")
}
colnames(df_FPR) <- c("Comparison", "Method", "FPR 0.01", "FPR 0.05",
"FPR 0.1")
df_FPR_melted <- reshape2::melt(df_FPR)
### Plot ###
# to avoid notes during the check
Method <- value <- variable <- y <- NULL
data_segm <- data.frame("y" = c(0.01, 0.05, 0.1),
"variable" = c("FPR 0.01", "FPR 0.05", "FPR 0.1"))
g <- ggplot(data = df_FPR_melted,
aes(color = Method,
x = tidytext::reorder_within(Method, value, variable,
fun = stats::median), y = value)) +
geom_boxplot() +
facet_grid(~ variable, scales = "free_x") +
tidytext::scale_x_reordered() +
geom_hline(data = data_segm,
aes(yintercept = y),
color = "red", lty = 2) +
xlab("Method") +
ylab(expression(Observed ~ alpha)) +
ggtitle(label = "False Positive Rate", subtitle = "by FPR level") +
scale_color_manual(values = cols) +
theme(legend.position = "none", axis.text.x = element_text(
angle = 90, hjust = 1, vjust = 0.5))
return(g)
} # END - function: plotFPR
#' @title plotFDR
#'
#' @export
#' @importFrom plyr ddply
#' @importFrom reshape2 melt
#' @importFrom tidytext reorder_within scale_x_reordered
#' @import ggplot2
#' @description
#' Draw the nominal false discovery rates for the 0.01, 0.05, and 0.1 levels.
#'
#' @param df_FDR a \code{data.frame} produced by the \code{\link{createTIEC}}
#' function, containing the FDR values.
#' @param cols named vector of colors.
#'
#' @return A ggplot object.
#'
#' @examples
#' # Load some data
#' data(ps_stool_16S)
#'
#' # Generate the patterns for 10 mock comparison for an experiment
#' # (N = 1000 is suggested)
#' mocks <- createMocks(nsamples = phyloseq::nsamples(ps_stool_16S), N = 10)
#' head(mocks)
#'
#' # Add some normalization/scaling factors to the phyloseq object
#' my_norm <- setNormalizations(fun = c("norm_edgeR", "norm_CSS"),
#' method = c("TMM", "CSS"))
#' ps_stool_16S <- runNormalizations(normalization_list = my_norm,
#' object = ps_stool_16S)
#'
#' # Initialize some limma based methods
#' my_limma <- set_limma(design = ~ group, coef = 2,
#' norm = c("TMM", "CSS"))
#'
#' # Run methods on mock datasets
#' results <- runMocks(mocks = mocks, method_list = my_limma,
#' object = ps_stool_16S)
#'
#' # Prepare results for Type I Error Control
#' TIEC_summary <- createTIEC(results)
#'
#' # Plot the results
#' plotFPR(df_FPR = TIEC_summary$df_FPR)
#' plotFDR(df_FDR = TIEC_summary$df_FDR)
#' plotQQ(df_QQ = TIEC_summary$df_QQ, zoom = c(0, 0.1))
#' plotKS(df_KS = TIEC_summary$df_KS)
#' plotLogP(df_QQ = TIEC_summary$df_QQ)
plotFDR <- function(df_FDR, cols = NULL) {
# Create the vector of colors
if (is.null(cols)) {
cols <- createColors(variable = df_FDR$Method)
}
if (is.null(names(cols))) {
stop("'cols' vector is not a named vector of colors.")
}
colnames(df_FDR) <- c("Method", "FDR 0.01", "FDR 0.05", "FDR 0.1")
df_FDR_melted <- reshape2::melt(df_FDR)
### Plot ###
# to avoid notes during the check
Method <- value <- variable <- y <- NULL
data_segm <- data.frame("y" = c(0.01, 0.05, 0.1),
"variable" = c("FDR 0.01", "FDR 0.05", "FDR 0.1"))
g <- ggplot(data = df_FDR_melted, aes(color = Method,
x = tidytext::reorder_within(Method, value, variable,
fun = stats::median),
y = value)) +
geom_point() +
facet_grid(~ variable, scales = "free_x") +
tidytext::scale_x_reordered() +
geom_hline(data = data_segm,
aes(yintercept = y),
color = "red", lty = 2) +
xlab("Method") +
ylab("Average FDR") +
ggtitle(label = "False Discovery Rate",
subtitle = "by significance thresholds") +
scale_color_manual(values = cols) +
theme(legend.position = "none", axis.text.x = element_text(
angle = 90, hjust = 1, vjust = 0.5))
return(g)
} # END - function: plotFDR
#' @title plotQQ
#'
#' @export
#' @importFrom plyr ddply
#' @import ggplot2
#' @description
#' Draw the average QQ-plots across the mock comparisons.
#'
#' @param df_QQ Coordinates to draw the QQ-plot to compare the mean observed
#' p-value distribution across comparisons, with the theoretical uniform
#' distribution.
#' @param cols named vector of colors.
#' @param zoom 2-dimesional vector containing the starting and the
#' final coordinates (default: \code{c(0, 0.1)})
#' @param split boolean value. If \code{TRUE}, the qq-plots are
#' reported separately for each method (default \code{split = FALSE}). Setting
#' it to \code{TRUE} is hardly suggested when the number of methods is high or
#' when their colors are similar.
#'
#' @return A ggplot object.
#'
#' @examples
#' # Load some data
#' data(ps_stool_16S)
#'
#' # Generate the patterns for 10 mock comparison for an experiment
#' # (N = 1000 is suggested)
#' mocks <- createMocks(nsamples = phyloseq::nsamples(ps_stool_16S), N = 10)
#' head(mocks)
#'
#' # Add some normalization/scaling factors to the phyloseq object
#' my_norm <- setNormalizations(fun = c("norm_edgeR", "norm_CSS"),
#' method = c("TMM", "CSS"))
#' ps_stool_16S <- runNormalizations(normalization_list = my_norm,
#' object = ps_stool_16S)
#'
#' # Initialize some limma based methods
#' my_limma <- set_limma(design = ~ group, coef = 2,
#' norm = c("TMM", "CSS"))
#'
#' # Run methods on mock datasets
#' results <- runMocks(mocks = mocks, method_list = my_limma,
#' object = ps_stool_16S)
#'
#' # Prepare results for Type I Error Control
#' TIEC_summary <- createTIEC(results)
#'
#' # Plot the results
#' plotFPR(df_FPR = TIEC_summary$df_FPR)
#' plotFDR(df_FDR = TIEC_summary$df_FDR)
#' plotQQ(df_QQ = TIEC_summary$df_QQ, zoom = c(0, 0.1))
#' plotKS(df_KS = TIEC_summary$df_KS)
#' plotLogP(df_QQ = TIEC_summary$df_QQ)
plotQQ <- function(df_QQ, cols = NULL, zoom = c(0, 0.1), split = FALSE) {
# Create the vector of colors
if (is.null(cols)) {
cols <- createColors(variable = df_QQ$Method)
}
if (is.null(names(cols))) {
stop("'cols' vector is not a named vector of colors.")
}
# Create segment for the 0 to 0.1 zoom of the plot
df_segments <- data.frame(x1 = c(0.01, 0, 0.05, 0, 0.1, 0), x2 = c(
0.01,
0.01, 0.05, 0.05, 0.1, 0.1
), y1 = c(0, 0.01, 0, 0.05, 0, 0.1), y2 = c(
0.01, 0.01, 0.05, 0.05, 0.1, 0.1
))
# to avoid notes during the check
Method <- pval_theoretical <- pval_observed <- x1 <- x2 <- y1 <- y2 <- NULL
g <- ggplot(data = df_QQ, aes(x = pval_theoretical, y = pval_observed,
color = Method)) +
geom_smooth(se = FALSE, na.rm = TRUE, span = 0.3, method = "loess",
formula = y ~ x) +
geom_abline() +
coord_equal(xlim = zoom, ylim = zoom) +
xlab("Theoretical p-value") +
ylab("Observed p-value") +
ggtitle(label = "Average QQ-plot", subtitle = paste(
"From",
zoom[1], "to", zoom[2]
)) +
geom_segment(aes(
x = x1, xend = x2, y = y1,
yend = y2
), color = "red", lty = 2, data = df_segments) +
scale_color_manual(values = cols)
if(split){
g <- g +
facet_wrap(~ Method) +
theme(legend.position = "none")
}
return(g)
}
#' @title plotLogP
#'
#' @export
#' @importFrom plyr ddply
#' @import ggplot2
#' @importFrom ggridges geom_density_ridges2 position_raincloud
#' @importFrom stats runif median p.adjust
#' @description
#' Draw the p-values or the average p-values distribution across the mock
#' comparisons in logarithmic scale.
#'
#' @param df_pval a \code{data.frame} produced by the \code{\link{createTIEC}}
#' function, containing the p-values for each taxon, method, and mock
#' comparison. It is used to draw the negative log10 p-values distribution.
#' If df_pval is supplied, let \code{df_QQ = NULL}.
#' @param df_QQ a \code{data.frame} produced by the \code{\link{createTIEC}}
#' function, containing the average p-values for each quantile and method. It is
#' used to draw the negative log10 average p-values distribution. If df_QQ is
#' supplied, let \code{df_pval = NULL}.
#' @param cols named vector of colors.
#'
#' @return A ggplot object.
#'
#' @examples
#' # Load some data
#' data(ps_stool_16S)
#'
#' # Generate the patterns for 10 mock comparison for an experiment
#' # (N = 1000 is suggested)
#' mocks <- createMocks(nsamples = phyloseq::nsamples(ps_stool_16S), N = 10)
#' head(mocks)
#'
#' # Add some normalization/scaling factors to the phyloseq object
#' my_norm <- setNormalizations(fun = c("norm_edgeR", "norm_CSS"),
#' method = c("TMM", "CSS"))
#' ps_stool_16S <- runNormalizations(normalization_list = my_norm,
#' object = ps_stool_16S)
#'
#' # Initialize some limma based methods
#' my_limma <- set_limma(design = ~ group, coef = 2,
#' norm = c("TMM", "CSS"))
#'
#' # Run methods on mock datasets
#' results <- runMocks(mocks = mocks, method_list = my_limma,
#' object = ps_stool_16S)
#'
#' # Prepare results for Type I Error Control
#' TIEC_summary <- createTIEC(results)
#'
#' # Plot the results
#' plotFPR(df_FPR = TIEC_summary$df_FPR)
#' plotFDR(df_FDR = TIEC_summary$df_FDR)
#' plotQQ(df_QQ = TIEC_summary$df_QQ, zoom = c(0, 0.1))
#' plotKS(df_KS = TIEC_summary$df_KS)
#' plotLogP(df_QQ = TIEC_summary$df_QQ)
plotLogP <- function(df_pval = NULL, df_QQ = NULL, cols = NULL) {
if(is.null(df_pval) & is.null(df_QQ)){
stop("Please supply 'df_pval' for all p-values analysis or ",
" 'df_QQ' for average p-values analysis.")
}
if(!is.null(df_pval) & !is.null(df_QQ)){
stop("Please supply one between 'df_pval' for all p-values analysis",
" or 'df_QQ' for average p-values analysis.")
}
# Add IDEAL distribution
p <- stats::runif(10000)
if(!is.null(df_pval)){
gtitle <- "All p-values log distribution"
df_pval_ideal <- data.frame(Comparison = "Comparison IDEAL",
Method = "IDEAL", variable = "IDEAL", pval = p,
padj = stats::p.adjust(p, method = "BH"))
df_to_plot <- rbind(df_pval, df_pval_ideal)
# Create the vector of colors
if (is.null(cols)) {
cols <- createColors(variable = df_pval$Method)
}
}
if(!is.null(df_QQ)){
gtitle <- "Average p-values log distribution"
df_pval_ideal <- data.frame(Method = "IDEAL",
pval_theoretical = p, pval_observed = p)
df_to_plot <- rbind(df_QQ, df_pval_ideal)
colnames(df_to_plot) <- c("Method", "pval_theoretical", "pval")
# Create the vector of colors
if (is.null(cols)) {
cols <- createColors(variable = df_QQ$Method)
}
}
if (is.null(names(cols))) {
stop("'cols' vector is not a named vector of colors.")
}
# Order method by closeness to IDEAL method
ord <- order(plyr::ddply(.data = df_to_plot, .variables = ~ Method, .fun =
function(x){
sum(stats::quantile(-log10(x$pval), c(0.99, 0.95, 0.9),
na.rm = TRUE) + log10(c(0.01, 0.05, 0.1)))
})$V1, decreasing = TRUE)
ordered_levels <- levels(df_to_plot$Method)[ord]
IDEAL_index <- which(ordered_levels == "IDEAL")
IDEAL_first <- c(ordered_levels[-IDEAL_index], "IDEAL")
df_to_plot$Method <- factor(df_to_plot$Method,
levels = IDEAL_first,
ordered = TRUE)
bold <- "IDEAL"
bold.labels <- ifelse(is.element(levels(df_to_plot[, "Method"]), bold),
yes = "bold", no = "plain")
color.labels <- ifelse(is.element(levels(df_to_plot[, "Method"]), bold),
yes = "red", no = "black")
# to avoid notes during the check
Method <- pval <- NULL
g <- ggplot(data = df_to_plot,
aes(x = -log10(pval), y = Method, fill = Method)) +
# Add histogram
ggridges::geom_density_ridges2(stat = "binline", binwidth = 0.1,
scale = 0.9, alpha = 0.5, draw_baseline = FALSE, na.rm = TRUE,
boundary = TRUE) +
# Add density
ggridges::geom_density_ridges(aes(vline_color = stat(quantile)),
alpha = 0.2, scale = 0.9, na.rm = TRUE, quantile_lines = TRUE,
quantiles = c(0.9, 0.95, 0.99), vline_size = 1, from = 0,
bandwidth = 0.1,
position = ggridges::position_raincloud(adjust_vlines = TRUE)) +
geom_vline(xintercept = -log10(c(0.1, 0.05, 0.01)), lty = 2) +
ylab("Method") +
xlab(expression(-log[10](p-value))) +
scale_x_continuous(sec.axis = sec_axis(trans = ~., name = "p-value",
breaks = -log10(c(0.1, 0.05, 0.01)), labels = c(0.1, 0.05, 0.01))) +
ggtitle(label = gtitle, subtitle =
paste0("IDEAL method as reference, red bars represents the 90,",
" 95, and 99 percentiles")) +
scale_fill_manual(values = c(cols, "IDEAL" = "#ff0000")) +
scale_color_manual(aesthetics = "vline_color",
values = c("#ffbaba", "#ff7b7b", "#ff5252", "#ff0000")) +
theme(legend.position = "none",
axis.text.y = suppressWarnings(element_text(face = bold.labels,
colour = color.labels, vjust = 0)),
axis.text.x.top = element_text(angle = 90, hjust = 0, vjust = 0.5))
return(g)
}
#' @title plotKS
#'
#' @export
#' @importFrom plyr ddply
#' @import ggplot2
#' @description
#' Draw the boxplots of the Kolmogorov-Smirnov test statistics for the p-value
#' distributions across the mock comparisons.
#'
#' @param df_KS a \code{data.frame} produced by the \code{\link{createTIEC}}
#' function containing the KS statistics and their p-values.
#' @param cols named vector of colors.
#'
#' @return A ggplot object.
#'
#' @examples
#' # Load some data
#' data(ps_stool_16S)
#'
#' # Generate the patterns for 10 mock comparison for an experiment
#' # (N = 1000 is suggested)
#' mocks <- createMocks(nsamples = phyloseq::nsamples(ps_stool_16S), N = 10)
#' head(mocks)
#'
#' # Add some normalization/scaling factors to the phyloseq object
#' my_norm <- setNormalizations(fun = c("norm_edgeR", "norm_CSS"),
#' method = c("TMM", "CSS"))
#' ps_stool_16S <- runNormalizations(normalization_list = my_norm,
#' object = ps_stool_16S)
#'
#' # Initialize some limma based methods
#' my_limma <- set_limma(design = ~ group, coef = 2,
#' norm = c("TMM", "CSS"))
#'
#' # Run methods on mock datasets
#' results <- runMocks(mocks = mocks, method_list = my_limma,
#' object = ps_stool_16S)
#'
#' # Prepare results for Type I Error Control
#' TIEC_summary <- createTIEC(results)
#'
#' # Plot the results
#' plotFPR(df_FPR = TIEC_summary$df_FPR)
#' plotFDR(df_FDR = TIEC_summary$df_FDR)
#' plotQQ(df_QQ = TIEC_summary$df_QQ, zoom = c(0, 0.1))
#' plotKS(df_KS = TIEC_summary$df_KS)
#' plotLogP(df_QQ = TIEC_summary$df_QQ)
plotKS <- function(df_KS, cols = NULL) {
# Create the vector of colors
if (is.null(cols)) {
cols <- createColors(variable = df_KS$Method)
}
if (is.null(names(cols))) {
stop("'cols' vector is not a named vector of colors.")
}
ord <- order(plyr::ddply(
.data = df_KS, .variables = ~ Method, .fun =
function(x) median(x$KS)
)$V1)
df_KS$Method <- factor(df_KS$Method,
levels = levels(df_KS$Method)[ord],
ordered = TRUE
)
# to avoid notes during the check
Method <- KS <- NULL
g <- ggplot(df_KS, aes(
x = Method, y = KS, color =
Method
)) +
geom_boxplot() +
coord_flip() +
ylab("KS statistics") +
xlab("Methods") +
ggtitle(
label = "K-S statistics", subtitle =
"Methods ordered by median K-S"
) +
scale_x_discrete(
limits = rev(levels(df_KS$Method)), position =
"top"
) +
theme(legend.position = "none", panel.spacing = unit(
0,
"lines"
), plot.margin = unit(c(0, 0, 0, 1), "cm")) +
scale_color_manual(values = cols)
return(g)
}
mcalgaro93/benchdamic documentation built on Nov. 28, 2024, 2:16 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 plotKS plotLogP plotQQ plotFDR plotFPR
Documented in plotFDR plotFPR plotKS plotLogP plotQQ
#' @title plotFPR
#'
#' @export
#' @importFrom plyr ddply
#' @importFrom reshape2 melt
#' @importFrom tidytext reorder_within scale_x_reordered
#' @import ggplot2
#' @description
#' Draw the boxplots of the proportions of p-values lower than 0.01, 0.05, and
#' 0.1 thresholds for each method.
#'
#' @param df_FPR a \code{data.frame} produced by the \code{\link{createTIEC}}
#' function, containing the FPR values.
#' @param cols named vector of colors.
#'
#' @return A ggplot object.
#'
#' @examples
#' # Load some data
#' data(ps_stool_16S)
#'
#' # Generate the patterns for 10 mock comparison for an experiment
#' # (N = 1000 is suggested)
#' mocks <- createMocks(nsamples = phyloseq::nsamples(ps_stool_16S), N = 10)
#' head(mocks)
#'
#' # Add some normalization/scaling factors to the phyloseq object
#' my_norm <- setNormalizations(fun = c("norm_edgeR", "norm_CSS"),
#' method = c("TMM", "CSS"))
#' ps_stool_16S <- runNormalizations(normalization_list = my_norm,
#' object = ps_stool_16S)
#'
#' # Initialize some limma based methods
#' my_limma <- set_limma(design = ~ group, coef = 2,
#' norm = c("TMM", "CSS"))
#'
#' # Run methods on mock datasets
#' results <- runMocks(mocks = mocks, method_list = my_limma,
#' object = ps_stool_16S)
#'
#' # Prepare results for Type I Error Control
#' TIEC_summary <- createTIEC(results)
#'
#' # Plot the results
#' plotFPR(df_FPR = TIEC_summary$df_FPR)
#' plotFDR(df_FDR = TIEC_summary$df_FDR)
#' plotQQ(df_QQ = TIEC_summary$df_QQ, zoom = c(0, 0.1))
#' plotKS(df_KS = TIEC_summary$df_KS)
#' plotLogP(df_QQ = TIEC_summary$df_QQ)
plotFPR <- function(df_FPR, cols = NULL) {
# Create the vector of colors
if (is.null(cols)) {
cols <- createColors(variable = df_FPR$Method)
}
if (is.null(names(cols))) {
stop("'cols' vector is not a named vector of colors.")
}
colnames(df_FPR) <- c("Comparison", "Method", "FPR 0.01", "FPR 0.05",
"FPR 0.1")
df_FPR_melted <- reshape2::melt(df_FPR)
### Plot ###
# to avoid notes during the check
Method <- value <- variable <- y <- NULL
data_segm <- data.frame("y" = c(0.01, 0.05, 0.1),
"variable" = c("FPR 0.01", "FPR 0.05", "FPR 0.1"))
g <- ggplot(data = df_FPR_melted,
aes(color = Method,
x = tidytext::reorder_within(Method, value, variable,
fun = stats::median), y = value)) +
geom_boxplot() +
facet_grid(~ variable, scales = "free_x") +
tidytext::scale_x_reordered() +
geom_hline(data = data_segm,
aes(yintercept = y),
color = "red", lty = 2) +
xlab("Method") +
ylab(expression(Observed ~ alpha)) +
ggtitle(label = "False Positive Rate", subtitle = "by FPR level") +
scale_color_manual(values = cols) +
theme(legend.position = "none", axis.text.x = element_text(
angle = 90, hjust = 1, vjust = 0.5))
return(g)
} # END - function: plotFPR
#' @title plotFDR
#'
#' @export
#' @importFrom plyr ddply
#' @importFrom reshape2 melt
#' @importFrom tidytext reorder_within scale_x_reordered
#' @import ggplot2
#' @description
#' Draw the nominal false discovery rates for the 0.01, 0.05, and 0.1 levels.
#'
#' @param df_FDR a \code{data.frame} produced by the \code{\link{createTIEC}}
#' function, containing the FDR values.
#' @param cols named vector of colors.
#'
#' @return A ggplot object.
#'
#' @examples
#' # Load some data
#' data(ps_stool_16S)
#'
#' # Generate the patterns for 10 mock comparison for an experiment
#' # (N = 1000 is suggested)
#' mocks <- createMocks(nsamples = phyloseq::nsamples(ps_stool_16S), N = 10)
#' head(mocks)
#'
#' # Add some normalization/scaling factors to the phyloseq object
#' my_norm <- setNormalizations(fun = c("norm_edgeR", "norm_CSS"),
#' method = c("TMM", "CSS"))
#' ps_stool_16S <- runNormalizations(normalization_list = my_norm,
#' object = ps_stool_16S)
#'
#' # Initialize some limma based methods
#' my_limma <- set_limma(design = ~ group, coef = 2,
#' norm = c("TMM", "CSS"))
#'
#' # Run methods on mock datasets
#' results <- runMocks(mocks = mocks, method_list = my_limma,
#' object = ps_stool_16S)
#'
#' # Prepare results for Type I Error Control
#' TIEC_summary <- createTIEC(results)
#'
#' # Plot the results
#' plotFPR(df_FPR = TIEC_summary$df_FPR)
#' plotFDR(df_FDR = TIEC_summary$df_FDR)
#' plotQQ(df_QQ = TIEC_summary$df_QQ, zoom = c(0, 0.1))
#' plotKS(df_KS = TIEC_summary$df_KS)
#' plotLogP(df_QQ = TIEC_summary$df_QQ)
plotFDR <- function(df_FDR, cols = NULL) {
# Create the vector of colors
if (is.null(cols)) {
cols <- createColors(variable = df_FDR$Method)
}
if (is.null(names(cols))) {
stop("'cols' vector is not a named vector of colors.")
}
colnames(df_FDR) <- c("Method", "FDR 0.01", "FDR 0.05", "FDR 0.1")
df_FDR_melted <- reshape2::melt(df_FDR)
### Plot ###
# to avoid notes during the check
Method <- value <- variable <- y <- NULL
data_segm <- data.frame("y" = c(0.01, 0.05, 0.1),
"variable" = c("FDR 0.01", "FDR 0.05", "FDR 0.1"))
g <- ggplot(data = df_FDR_melted, aes(color = Method,
x = tidytext::reorder_within(Method, value, variable,
fun = stats::median),
y = value)) +
geom_point() +
facet_grid(~ variable, scales = "free_x") +
tidytext::scale_x_reordered() +
geom_hline(data = data_segm,
aes(yintercept = y),
color = "red", lty = 2) +
xlab("Method") +
ylab("Average FDR") +
ggtitle(label = "False Discovery Rate",
subtitle = "by significance thresholds") +
scale_color_manual(values = cols) +
theme(legend.position = "none", axis.text.x = element_text(
angle = 90, hjust = 1, vjust = 0.5))
return(g)
} # END - function: plotFDR
#' @title plotQQ
#'
#' @export
#' @importFrom plyr ddply
#' @import ggplot2
#' @description
#' Draw the average QQ-plots across the mock comparisons.
#'
#' @param df_QQ Coordinates to draw the QQ-plot to compare the mean observed
#' p-value distribution across comparisons, with the theoretical uniform
#' distribution.
#' @param cols named vector of colors.
#' @param zoom 2-dimesional vector containing the starting and the
#' final coordinates (default: \code{c(0, 0.1)})
#' @param split boolean value. If \code{TRUE}, the qq-plots are
#' reported separately for each method (default \code{split = FALSE}). Setting
#' it to \code{TRUE} is hardly suggested when the number of methods is high or
#' when their colors are similar.
#'
#' @return A ggplot object.
#'
#' @examples
#' # Load some data
#' data(ps_stool_16S)
#'
#' # Generate the patterns for 10 mock comparison for an experiment
#' # (N = 1000 is suggested)
#' mocks <- createMocks(nsamples = phyloseq::nsamples(ps_stool_16S), N = 10)
#' head(mocks)
#'
#' # Add some normalization/scaling factors to the phyloseq object
#' my_norm <- setNormalizations(fun = c("norm_edgeR", "norm_CSS"),
#' method = c("TMM", "CSS"))
#' ps_stool_16S <- runNormalizations(normalization_list = my_norm,
#' object = ps_stool_16S)
#'
#' # Initialize some limma based methods
#' my_limma <- set_limma(design = ~ group, coef = 2,
#' norm = c("TMM", "CSS"))
#'
#' # Run methods on mock datasets
#' results <- runMocks(mocks = mocks, method_list = my_limma,
#' object = ps_stool_16S)
#'
#' # Prepare results for Type I Error Control
#' TIEC_summary <- createTIEC(results)
#'
#' # Plot the results
#' plotFPR(df_FPR = TIEC_summary$df_FPR)
#' plotFDR(df_FDR = TIEC_summary$df_FDR)
#' plotQQ(df_QQ = TIEC_summary$df_QQ, zoom = c(0, 0.1))
#' plotKS(df_KS = TIEC_summary$df_KS)
#' plotLogP(df_QQ = TIEC_summary$df_QQ)
plotQQ <- function(df_QQ, cols = NULL, zoom = c(0, 0.1), split = FALSE) {
# Create the vector of colors
if (is.null(cols)) {
cols <- createColors(variable = df_QQ$Method)
}
if (is.null(names(cols))) {
stop("'cols' vector is not a named vector of colors.")
}
# Create segment for the 0 to 0.1 zoom of the plot
df_segments <- data.frame(x1 = c(0.01, 0, 0.05, 0, 0.1, 0), x2 = c(
0.01,
0.01, 0.05, 0.05, 0.1, 0.1
), y1 = c(0, 0.01, 0, 0.05, 0, 0.1), y2 = c(
0.01, 0.01, 0.05, 0.05, 0.1, 0.1
))
# to avoid notes during the check
Method <- pval_theoretical <- pval_observed <- x1 <- x2 <- y1 <- y2 <- NULL
g <- ggplot(data = df_QQ, aes(x = pval_theoretical, y = pval_observed,
color = Method)) +
geom_smooth(se = FALSE, na.rm = TRUE, span = 0.3, method = "loess",
formula = y ~ x) +
geom_abline() +
coord_equal(xlim = zoom, ylim = zoom) +
xlab("Theoretical p-value") +
ylab("Observed p-value") +
ggtitle(label = "Average QQ-plot", subtitle = paste(
"From",
zoom[1], "to", zoom[2]
)) +
geom_segment(aes(
x = x1, xend = x2, y = y1,
yend = y2
), color = "red", lty = 2, data = df_segments) +
scale_color_manual(values = cols)
if(split){
g <- g +
facet_wrap(~ Method) +
theme(legend.position = "none")
}
return(g)
}
#' @title plotLogP
#'
#' @export
#' @importFrom plyr ddply
#' @import ggplot2
#' @importFrom ggridges geom_density_ridges2 position_raincloud
#' @importFrom stats runif median p.adjust
#' @description
#' Draw the p-values or the average p-values distribution across the mock
#' comparisons in logarithmic scale.
#'
#' @param df_pval a \code{data.frame} produced by the \code{\link{createTIEC}}
#' function, containing the p-values for each taxon, method, and mock
#' comparison. It is used to draw the negative log10 p-values distribution.
#' If df_pval is supplied, let \code{df_QQ = NULL}.
#' @param df_QQ a \code{data.frame} produced by the \code{\link{createTIEC}}
#' function, containing the average p-values for each quantile and method. It is
#' used to draw the negative log10 average p-values distribution. If df_QQ is
#' supplied, let \code{df_pval = NULL}.
#' @param cols named vector of colors.
#'
#' @return A ggplot object.
#'
#' @examples
#' # Load some data
#' data(ps_stool_16S)
#'
#' # Generate the patterns for 10 mock comparison for an experiment
#' # (N = 1000 is suggested)
#' mocks <- createMocks(nsamples = phyloseq::nsamples(ps_stool_16S), N = 10)
#' head(mocks)
#'
#' # Add some normalization/scaling factors to the phyloseq object
#' my_norm <- setNormalizations(fun = c("norm_edgeR", "norm_CSS"),
#' method = c("TMM", "CSS"))
#' ps_stool_16S <- runNormalizations(normalization_list = my_norm,
#' object = ps_stool_16S)
#'
#' # Initialize some limma based methods
#' my_limma <- set_limma(design = ~ group, coef = 2,
#' norm = c("TMM", "CSS"))
#'
#' # Run methods on mock datasets
#' results <- runMocks(mocks = mocks, method_list = my_limma,
#' object = ps_stool_16S)
#'
#' # Prepare results for Type I Error Control
#' TIEC_summary <- createTIEC(results)
#'
#' # Plot the results
#' plotFPR(df_FPR = TIEC_summary$df_FPR)
#' plotFDR(df_FDR = TIEC_summary$df_FDR)
#' plotQQ(df_QQ = TIEC_summary$df_QQ, zoom = c(0, 0.1))
#' plotKS(df_KS = TIEC_summary$df_KS)
#' plotLogP(df_QQ = TIEC_summary$df_QQ)
plotLogP <- function(df_pval = NULL, df_QQ = NULL, cols = NULL) {
if(is.null(df_pval) & is.null(df_QQ)){
stop("Please supply 'df_pval' for all p-values analysis or ",
" 'df_QQ' for average p-values analysis.")
}
if(!is.null(df_pval) & !is.null(df_QQ)){
stop("Please supply one between 'df_pval' for all p-values analysis",
" or 'df_QQ' for average p-values analysis.")
}
# Add IDEAL distribution
p <- stats::runif(10000)
if(!is.null(df_pval)){
gtitle <- "All p-values log distribution"
df_pval_ideal <- data.frame(Comparison = "Comparison IDEAL",
Method = "IDEAL", variable = "IDEAL", pval = p,
padj = stats::p.adjust(p, method = "BH"))
df_to_plot <- rbind(df_pval, df_pval_ideal)
# Create the vector of colors
if (is.null(cols)) {
cols <- createColors(variable = df_pval$Method)
}
}
if(!is.null(df_QQ)){
gtitle <- "Average p-values log distribution"
df_pval_ideal <- data.frame(Method = "IDEAL",
pval_theoretical = p, pval_observed = p)
df_to_plot <- rbind(df_QQ, df_pval_ideal)
colnames(df_to_plot) <- c("Method", "pval_theoretical", "pval")
# Create the vector of colors
if (is.null(cols)) {
cols <- createColors(variable = df_QQ$Method)
}
}
if (is.null(names(cols))) {
stop("'cols' vector is not a named vector of colors.")
}
# Order method by closeness to IDEAL method
ord <- order(plyr::ddply(.data = df_to_plot, .variables = ~ Method, .fun =
function(x){
sum(stats::quantile(-log10(x$pval), c(0.99, 0.95, 0.9),
na.rm = TRUE) + log10(c(0.01, 0.05, 0.1)))
})$V1, decreasing = TRUE)
ordered_levels <- levels(df_to_plot$Method)[ord]
IDEAL_index <- which(ordered_levels == "IDEAL")
IDEAL_first <- c(ordered_levels[-IDEAL_index], "IDEAL")
df_to_plot$Method <- factor(df_to_plot$Method,
levels = IDEAL_first,
ordered = TRUE)
bold <- "IDEAL"
bold.labels <- ifelse(is.element(levels(df_to_plot[, "Method"]), bold),
yes = "bold", no = "plain")
color.labels <- ifelse(is.element(levels(df_to_plot[, "Method"]), bold),
yes = "red", no = "black")
# to avoid notes during the check
Method <- pval <- NULL
g <- ggplot(data = df_to_plot,
aes(x = -log10(pval), y = Method, fill = Method)) +
# Add histogram
ggridges::geom_density_ridges2(stat = "binline", binwidth = 0.1,
scale = 0.9, alpha = 0.5, draw_baseline = FALSE, na.rm = TRUE,
boundary = TRUE) +
# Add density
ggridges::geom_density_ridges(aes(vline_color = stat(quantile)),
alpha = 0.2, scale = 0.9, na.rm = TRUE, quantile_lines = TRUE,
quantiles = c(0.9, 0.95, 0.99), vline_size = 1, from = 0,
bandwidth = 0.1,
position = ggridges::position_raincloud(adjust_vlines = TRUE)) +
geom_vline(xintercept = -log10(c(0.1, 0.05, 0.01)), lty = 2) +
ylab("Method") +
xlab(expression(-log[10](p-value))) +
scale_x_continuous(sec.axis = sec_axis(trans = ~., name = "p-value",
breaks = -log10(c(0.1, 0.05, 0.01)), labels = c(0.1, 0.05, 0.01))) +
ggtitle(label = gtitle, subtitle =
paste0("IDEAL method as reference, red bars represents the 90,",
" 95, and 99 percentiles")) +
scale_fill_manual(values = c(cols, "IDEAL" = "#ff0000")) +
scale_color_manual(aesthetics = "vline_color",
values = c("#ffbaba", "#ff7b7b", "#ff5252", "#ff0000")) +
theme(legend.position = "none",
axis.text.y = suppressWarnings(element_text(face = bold.labels,
colour = color.labels, vjust = 0)),
axis.text.x.top = element_text(angle = 90, hjust = 0, vjust = 0.5))
return(g)
}
#' @title plotKS
#'
#' @export
#' @importFrom plyr ddply
#' @import ggplot2
#' @description
#' Draw the boxplots of the Kolmogorov-Smirnov test statistics for the p-value
#' distributions across the mock comparisons.
#'
#' @param df_KS a \code{data.frame} produced by the \code{\link{createTIEC}}
#' function containing the KS statistics and their p-values.
#' @param cols named vector of colors.
#'
#' @return A ggplot object.
#'
#' @examples
#' # Load some data
#' data(ps_stool_16S)
#'
#' # Generate the patterns for 10 mock comparison for an experiment
#' # (N = 1000 is suggested)
#' mocks <- createMocks(nsamples = phyloseq::nsamples(ps_stool_16S), N = 10)
#' head(mocks)
#'
#' # Add some normalization/scaling factors to the phyloseq object
#' my_norm <- setNormalizations(fun = c("norm_edgeR", "norm_CSS"),
#' method = c("TMM", "CSS"))
#' ps_stool_16S <- runNormalizations(normalization_list = my_norm,
#' object = ps_stool_16S)
#'
#' # Initialize some limma based methods
#' my_limma <- set_limma(design = ~ group, coef = 2,
#' norm = c("TMM", "CSS"))
#'
#' # Run methods on mock datasets
#' results <- runMocks(mocks = mocks, method_list = my_limma,
#' object = ps_stool_16S)
#'
#' # Prepare results for Type I Error Control
#' TIEC_summary <- createTIEC(results)
#'
#' # Plot the results
#' plotFPR(df_FPR = TIEC_summary$df_FPR)
#' plotFDR(df_FDR = TIEC_summary$df_FDR)
#' plotQQ(df_QQ = TIEC_summary$df_QQ, zoom = c(0, 0.1))
#' plotKS(df_KS = TIEC_summary$df_KS)
#' plotLogP(df_QQ = TIEC_summary$df_QQ)
plotKS <- function(df_KS, cols = NULL) {
# Create the vector of colors
if (is.null(cols)) {
cols <- createColors(variable = df_KS$Method)
}
if (is.null(names(cols))) {
stop("'cols' vector is not a named vector of colors.")
}
ord <- order(plyr::ddply(
.data = df_KS, .variables = ~ Method, .fun =
function(x) median(x$KS)
)$V1)
df_KS$Method <- factor(df_KS$Method,
levels = levels(df_KS$Method)[ord],
ordered = TRUE
)
# to avoid notes during the check
Method <- KS <- NULL
g <- ggplot(df_KS, aes(
x = Method, y = KS, color =
Method
)) +
geom_boxplot() +
coord_flip() +
ylab("KS statistics") +
xlab("Methods") +
ggtitle(
label = "K-S statistics", subtitle =
"Methods ordered by median K-S"
) +
scale_x_discrete(
limits = rev(levels(df_KS$Method)), position =
"top"
) +
theme(legend.position = "none", panel.spacing = unit(
0,
"lines"
), plot.margin = unit(c(0, 0, 0, 1), "cm")) +
scale_color_manual(values = cols)
return(g)
}
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.