suprarai commented 1 month ago

Hi, I want to change some of the parameters on the graph like changing x and y axis labels, adding the line on the X and Y axis.

I try to do it on this step but could not able to make it.

plot otc curves

OTC::plot_otc_curves(otc_curves = otc_curves, output_path ="../results", output_name = "demo")

Could you please help me if it is possible to do it ? Best, Surya

jn-goe commented 1 month ago

Hello Surya,

Thanks for your message! The plot_otc_curves function in the OTC package currently does not allow plot customizations. For now, you can try the following function plot_otc_curves_flexible, which I adapted to rename the x and y axis labels (function parameters name_xlab and name_ylab). The function now outputs the ggplot2 object itself, which can be saved and customized more flexibly with ggplot2 functions. To use the function, copy it and run it in your R session:

plot_otc_curves_flexible <- function(otc_curves, 
                                     load_data = FALSE, 
                                     CB = TRUE, 
                                     alpha = 0.05, 
                                     upper_t = 250, 
                                     lower_t = 30,
                                     name_xlab = "threshold in nm",
                                     name_ylab = "Optimal Transport Colocalization"){
  if (load_data) {
    env <- new.env()
    nm <- load(otc_curves, env)[1]
    otc_curves <- env[[nm]]
  }
  otc_means <- aggregate(otc_curves$OTC, by = list(otc_curves$class, 
                                                   otc_curves$t), mean)
  colnames(otc_means) <- c("class", "t", "OTC")
  if (CB) {
    cb_name = paste("CB_", toString(alpha), "_", sep = "")
    otc_means$quantile = 0
    for (class in unique(otc_curves$class)) {
      data <- otc_curves[otc_curves$class == class, ]
      n <- length(unique(data$no_pic))
      Ts <- unique(data$t)
      p <- length(Ts)
      df <- data.frame(matrix(ncol = p, nrow = n))
      colnames(df) <- Ts
      for (j in 1:n) {
        df[j, ] <- data[data$no_pic == j, ]$OTC
      }
      df_center <- scale(df, scale = FALSE)
      cov_matrix = matrix(0, p, p)
      for (i in 1:nrow(df_center)) {
        cov_matrix = cov_matrix + as.numeric(df_center[i, 
        ]) %*% t(as.numeric(df_center[i, ]))
      }
      cov_matrix = 1/(n - 1) * cov_matrix
      sample = MASS::mvrnorm(n = 1000, mu = rep(0, p), 
                             Sigma = cov_matrix)
      sample = abs(sample)
      sups = apply(sample, 1, max)
      alpha = 0.05
      u_alpha = quantile(sups, probs = alpha/2)
      otc_means[otc_means$class == class, ]$quantile <- u_alpha
    }
    otc_means$upper <- otc_means$OTC + otc_means$quantile/sqrt(n)
    otc_means$lower <- otc_means$OTC - otc_means$quantile/sqrt(n)
    otc_means$upper_thresh <- 1
    otc_means$lower_thresh <- 0
    otc_means$upper <- apply(otc_means[, c("upper", "upper_thresh")], 
                             1, min)
    otc_means$lower <- apply(otc_means[, c("lower", "lower_thresh")], 
                             1, max)
    otc_means <- subset(otc_means, select = -c(lower_thresh, 
                                               upper_thresh))
  }
  else {
    cb_name = "_"
  }
  otc_means <- otc_means[otc_means$t >= lower_t & otc_means$t <= 
                           upper_t, ]

  p <- ggplot(data = otc_means, aes(x = t, y = OTC, color = class)) + 
    geom_line() + 
    geom_ribbon(aes(x = t, ymax = upper, ymin = lower, fill = class), 
                alpha = 0.3) + 
    scale_x_continuous(name = name_xlab) + 
    scale_y_continuous(name = name_ylab, lim = c(0, 1)) + 
    guides(color = guide_legend(title = ""), 
           linetype = guide_legend(title = ""), 
           fill = guide_legend(title = "")) + 
    theme(legend.key.height = unit(2, "lines"), 
          legend.key.width = unit(2, "lines"))

  return(p)
}

For example, if you want to add a vertical line at 50 on the x-axis, you can do this as follows:

p <- plot_otc_curves_flexible(otc_curves = otc_curves, 
                         name_ylab = "new_ylab", 
                         name_xlab = "new_xlab")
p + ggplot2::geom_vline(xintercept = 50)

Let me know if that helps :)

Best, Julia

suprarai commented 2 weeks ago

Dear Julia, Thank you very much for the code. It means a lot to me.

I found a easier way. I first saved csv file and then recreated 95% OTC confidence band using ggplot2 and stat_summary function. This way, I can plot multiple OTC bands for multiple objects (lets say A, B, C, and D) with another object (lets say X) in a same figure.

But, I have another issue. I tried stat_summary function to make 95% confidence band. However, the width of the band (upper and lower threshold) is "bigger" as compared to the one created using OTC::plot_otc_curves. I guess is you might have used different formula to calculate the upper and lower threshold for the confidence band.

Could you please help me so that I can recreate exact similar plot from csv file as well ?. The csv file only spits "t" and "OTC" values. Thanks in advance! Best, Surya

ctameling / OTC

Manipulating ggplot2 parameters in otc_curve! #2

plot otc curves