Skip to content

Instantly share code, notes, and snippets.

@mbarnkob

mbarnkob/2016-Hazard-ratio-in-ggplot2

Created October 31, 2016 11:05
Show Gist options

  • Select an option

    Learn more about clone URLs
  • Save mbarnkob/c0fa8aae4c0347484248eedf90d493be to your computer and use it in GitHub Desktop.

Select an option

Learn more about clone URLs
Save mbarnkob/c0fa8aae4c0347484248eedf90d493be to your computer and use it in GitHub Desktop.
Download ZIP

Revisions

  1. mbarnkob created this gist Oct 31, 2016.
    33 changes: 33 additions & 0 deletions 2016-Hazard-ratio-in-ggplot2
    View file
    Original file line number Diff line number Diff line change
    @@ -0,0 +1,33 @@
    ### Visualization of hazard ratio's in TCGA data based on a single gene
    ### =================

    # Goal: To visualize survival data (deaths) based on the expression level of a single gene
    # as either high or low, and the hazard ratio between those two.

    # Date: 2016-10-30
    # Author: Mike Barnkob, www.mikebarnkob.dk

    ### Data

    df_data <- data.frame(Cancer=c("Brain", "Colorectal", "Kidney clear cell carcinoma", "Kidney renal papillary carcinoma"),
    HR=c(1.03, 0.98, 1.27, 1.22),
    HR_lower=c(0.97, 0.62, 1.16, 1.03),
    HR_upper=c(1.09, 1.55, 1.38, 1.45)
    )

    ### Visualize

    if (!require('ggplot2')) install.packages('ggplot2'); library('ggplot2') # Load ggplot2 library

    #Based on http://www.gettinggeneticsdone.com/2011/03/forest-plots-using-rstats-and-ggplot2.html and https://github.com/hadley/ggplot2/issues/1441
    #Layout inspired by Fivethirtyeight's election tables: http://projects.fivethirtyeight.com/2016-election-forecast/#margins

    p <- ggplot(df_data, aes(x=Cancer, y=HR, ymin=HR_lower, ymax=HR_upper)) +
    geom_linerange(size=8, colour="#a6d8f0") +
    geom_hline(aes(x=0, yintercept=1), lty=1) +
    geom_point(size=3, shape=21, fill="#008fd5", colour = "white", stroke = 1) +
    scale_y_continuous(limits = c(0.5, 2)) +
    coord_flip() +
    ggtitle("Hazard Ratio for Gene of Interest") +
    theme_minimal()
    p