plot_con_permute_all: plot_con_permute_all

In LyonsZhang/ProgPerm: Progressive permutation for a dynamic representation of the robustness of microbiome discoveries

Usage

plot_con_permute_all(alloutputs, top_pm, lgndcol = 2, legend = FALSE, psigtitle, psigyrange = NULL, savepsigfile, psigpicdim, pvtitle, pvyrange = NULL, savepvfile, pvpicdim, estitle, esyrange = NULL, saveesfile, espicdim, crtitle, cryrange = NULL, savecrfile, crpicdim)

Arguments

alloutputs
top_pm
lgndcol
legend
psigtitle
psigyrange
savepsigfile
psigpicdim
pvtitle
pvyrange
savepvfile
pvpicdim
estitle
esyrange
saveesfile
espicdim
crtitle
cryrange
savecrfile
crpicdim

Examples

##---- Should be DIRECTLY executable !! ----
##-- ==>  Define data, use random,
##--	or do  help(data=index)  for the standard data sets.

## The function is currently defined as
function (alloutputs, top_pm, lgndcol = 2, legend = FALSE, psigtitle, 
    psigyrange = NULL, savepsigfile, psigpicdim, pvtitle, pvyrange = NULL, 
    savepvfile, pvpicdim, estitle, esyrange = NULL, saveesfile, 
    espicdim, crtitle, cryrange = NULL, savecrfile, crpicdim) 
{
    require(ggplot2)
    results <- alloutputs$alldist
    selname0 <- alloutputs$selname
    if (top_pm > length(selname0)) {
        stop("top_pm should be smaller than the number of observations.")
    }
    df <- as.data.frame(results$psigdist)
    df$highlight <- ifelse(df$dif == 0, "real data", "permutation")
    df$highlight <- factor(df$highlight, levels = c("real data", 
        "permutation"))
    mycolours <- c(permutation = "black", `real data` = "red")
    N0 <- max(df$dif)
    if (sum(is.na(df))) {
        stop("Missing values exist in psigdist")
    }
    else {
        df$dif <- df$dif/N0
        if (is.null(psigyrange)) {
            y_range <- range(floor(min(df$q1)), ceiling(max(df$q2)))
        }
        else {
            y_range <- psigyrange
        }
        ggplot(df, aes(x = dif, y = median)) + geom_point(size = 5, 
            aes(colour = highlight)) + scale_color_manual("Scenarios", 
            values = mycolours) + geom_pointrange(aes(ymin = q1, 
            ymax = q2, colour = highlight), size = 0.5) + geom_line(size = 1) + 
            xlab("Proportion of permutations") + ylab("Number of significant features") + 
            ggtitle(psigtitle) + ylim(y_range) + theme_bw() + 
            theme(plot.title = element_text(size = 30, face = "bold"), 
                legend.title = element_text(size = 20), legend.text = element_text(size = 16)) + 
            theme(axis.line = element_line(colour = "black"), 
                panel.grid.major = element_blank(), panel.border = element_blank(), 
                panel.background = element_blank(), axis.line.x = element_line(color = "black", 
                  size = 1), axis.line.y = element_line(color = "black", 
                  size = 1), axis.text = element_text(size = 16, 
                  face = "bold"), axis.title = element_text(size = 20, 
                  face = "bold"))
        ggsave(filename = savepsigfile, width = psigpicdim[1], 
            height = psigpicdim[2])
    }
    df <- as.data.frame(results$pvdist)
    N <- length(unique(df$dif))
    df$dif <- factor(df$dif, levels = sort(unique(df$dif)))
    selname <- selname0[1:top_pm]
    if (top_pm < length(selname0)) {
        tt <- split(df, df$dif)
        ss <- lapply(tt, function(x) {
            return(x[1:top_pm, ])
        })
        df <- do.call("rbind", ss)
    }
    if (sum(is.na(df))) {
        stop("Missing values exist in pvdist")
    }
    else {
        df$dif <- as.numeric(as.character(df[, "dif"]))/N0
        df$median <- as.numeric(as.character(df[, "median"]))
        df$q1 <- as.numeric(as.character(df[, "q1"]))
        df$q2 <- as.numeric(as.character(df[, "q2"]))
        df$variable <- factor(df$variable, levels = selname)
        if (is.null(pvyrange)) {
            y_range <- range(floor(min(df$q1)), ceiling(max(df$q2)))
        }
        else {
            y_range <- pvyrange
        }
        p <- ggplot(data = df, aes(x = dif, y = median, colour = variable, 
            group = variable)) + geom_line() + geom_point() + 
            geom_pointrange(aes(ymin = q1, ymax = q2)) + ylim(y_range) + 
            xlab("Proportion of permutations") + ylab(paste("P-values of top", 
            paste(top_pm, "important features", sep = " "), sep = " ")) + 
            ggtitle(pvtitle) + theme_bw() + theme(plot.title = element_text(size = 30, 
            face = "bold"), axis.line = element_line(colour = "black"), 
            panel.grid.major = element_blank(), panel.border = element_blank(), 
            panel.background = element_blank(), axis.line.x = element_line(color = "black", 
                size = 1), axis.line.y = element_line(color = "black", 
                size = 1), axis.text = element_text(size = 16, 
                face = "bold"), axis.title = element_text(size = 20, 
                face = "bold"))
        if (!legend) {
            p <- p + theme(legend.position = "none")
        }
        else {
            p <- p + guides(col = guide_legend(ncol = lgndcol)) + 
                theme(legend.title = element_text(size = 14), 
                  legend.text = element_text(size = 10))
        }
        ggsave(filename = savepvfile, plot = p, width = pvpicdim[1], 
            height = pvpicdim[2])
    }
    df <- as.data.frame(results$esdist)
    N <- length(unique(df$dif))
    df$dif <- factor(df$dif, levels = sort(unique(df$dif)))
    selname <- selname0[1:top_pm]
    if (top_pm < length(selname0)) {
        tt <- split(df, df$dif)
        ss <- lapply(tt, function(x) {
            return(x[1:top_pm, ])
        })
        df <- do.call("rbind", ss)
    }
    if (sum(is.na(df))) {
        stop("Missing values exist in esdist")
    }
    else {
        df$dif <- as.numeric(as.character(df[, "dif"]))/N
        df$median <- as.numeric(as.character(df[, "median"]))
        df$q1 <- as.numeric(as.character(df[, "q1"]))
        df$q2 <- as.numeric(as.character(df[, "q2"]))
        df$variable <- factor(df$variable, levels = selname)
        if (is.null(esyrange)) {
            y_range <- range(min(df$q1) - 0.1, max(df$q2) + 0.1)
        }
        else {
            y_range <- esyrange
        }
        ggplot(data = df, aes(x = dif, y = median, colour = variable, 
            group = variable)) + geom_line() + geom_point() + 
            geom_pointrange(aes(ymin = q1, ymax = q2)) + ylim(y_range) + 
            xlab("Proportion of permutations") + ylab(paste("Effect sizes of top", 
            paste(top_pm, "important features", sep = " "), sep = " ")) + 
            ggtitle(estitle) + theme_bw() + theme(plot.title = element_text(size = 30, 
            face = "bold"), axis.line = element_line(colour = "black"), 
            panel.grid.major = element_blank(), panel.border = element_blank(), 
            panel.background = element_blank(), axis.line.x = element_line(color = "black", 
                size = 1), axis.line.y = element_line(color = "black", 
                size = 1), axis.text = element_text(size = 16, 
                face = "bold"), axis.title = element_text(size = 20, 
                face = "bold"))
        if (!legend) {
            p <- p + theme(legend.position = "none")
        }
        else {
            p <- p + guides(col = guide_legend(ncol = lgndcol)) + 
                theme(legend.title = element_text(size = 14), 
                  legend.text = element_text(size = 10))
        }
        ggsave(filename = saveesfile, plot = p, width = espicdim[1], 
            height = espicdim[2])
    }
    df <- as.data.frame(results$crdist)
    N <- length(unique(df$dif))
    df$dif <- factor(df$dif, levels = sort(unique(df$dif)))
    selname <- selname0[1:top_pm]
    if (top_pm < length(selname0)) {
        tt <- split(df, df$dif)
        ss <- lapply(tt, function(x) {
            return(x[1:top_pm, ])
        })
        df <- do.call("rbind", ss)
    }
    if (sum(is.na(df))) {
        stop("Missing values exist in crdist")
    }
    else {
        df$dif <- as.numeric(as.character(df[, "dif"]))/N0
        df$median <- as.numeric(as.character(df[, "median"]))
        df$q1 <- as.numeric(as.character(df[, "q1"]))
        df$q2 <- as.numeric(as.character(df[, "q2"]))
        df$variable <- factor(df$variable, levels = selname)
        if (is.null(cryrange)) {
            y_range <- range(min(df$q1) - 0.1, max(df$q2) + 0.1)
        }
        else {
            y_range <- cryrange
        }
        ggplot(data = df, aes(x = dif, y = median, colour = variable, 
            group = variable)) + geom_line() + geom_point() + 
            geom_pointrange(aes(ymin = q1, ymax = q2)) + ylim(y_range) + 
            xlab("Proportion of permutations") + ylab(paste("Correlations of top", 
            paste(top_pm, "important features", sep = " "), sep = " ")) + 
            ggtitle(crtitle) + theme_bw() + theme(plot.title = element_text(size = 30, 
            face = "bold"), axis.line = element_line(colour = "black"), 
            panel.grid.major = element_blank(), panel.border = element_blank(), 
            panel.background = element_blank(), axis.line.x = element_line(color = "black", 
                size = 1), axis.line.y = element_line(color = "black", 
                size = 1), axis.text = element_text(size = 16, 
                face = "bold"), axis.title = element_text(size = 20, 
                face = "bold"))
        if (!legend) {
            p <- p + theme(legend.position = "none")
        }
        else {
            p <- p + guides(col = guide_legend(ncol = lgndcol)) + 
                theme(legend.title = element_text(size = 14), 
                  legend.text = element_text(size = 10))
        }
        ggsave(filename = savecrfile, plot = p, width = crpicdim[1], 
            height = crpicdim[2])
    }
  }

LyonsZhang/ProgPerm documentation built on July 16, 2020, 12:45 a.m.