Libraries

library(tidyverse)
library(reshape2)
library(ggpubr)
library(lubridate)

Data

chim.df = read.csv("./data/aml.all.df.csv")
names(chim.df)
##  [1] "ID"        "dot"       "dor"       "txage"     "relapse"   "sex"      
##  [7] "rstatprtx" "ghgp"      "tbi"       "agvhd"     "test"      "bdate"    
## [13] "btime"     "rtime"     "rbin"      "bmc_cdw"   "bmc_cd3"   "bmc_cd15" 
## [19] "bmc_cd34"  "pdate"     "ptime"     "pbc_cdw"   "pbc_cd3"   "pbc_cd15" 
## [25] "pbc_cd34"
chim.df$ID = as.factor(chim.df$ID)

Remove plot relapse samples

plot.df = chim.df[which(chim.df$pdate < chim.df$dor | is.na(chim.df$dor)), ]

CDW proportions

  • Relapse vs. non-relapse
ggboxplot(plot.df, x = "relapse", y = "pbc_cdw", 
          color = "relapse", add = "jitter", main = "PB CD (Whole)") +
  stat_compare_means(label.y = 20, method = "wilcox.test") + 
  scale_y_continuous(limits = c(0,100))

  • Time plots
ggscatter(plot.df, x = "ptime", y = "pbc_cdw", col = "relapse", main = "PB CD (Whole)",
          add = "reg.line")

p1 = ggline(plot.df, x = "ptime", y = "pbc_cdw", col = "ID", 
       facet.by = "relapse", numeric.x.axis = TRUE, main = "PB CD (Whole)")
ggpar(p1, legend = "none")

  • Boxplots by testing period
ggboxplot(plot.df, x = "test", y = "pbc_cdw", fill = "relapse", main = "PB CD (Whole)")

CD3 proportions

  • Relapse vs. non-relapse
ggboxplot(plot.df, x = "relapse", y = "pbc_cd3", 
          color = "relapse", add = "jitter", main = "PB CD3") +
  stat_compare_means(label.y = 20, method = "wilcox.test") + 
  scale_y_continuous(limits = c(0,100))

  • Time plots
ggscatter(plot.df, x = "ptime", y = "pbc_cd3", col = "relapse", main = "PB CD3",
          add = "reg.line")

p1 = ggline(plot.df, x = "ptime", y = "pbc_cd3", col = "ID", 
       facet.by = "relapse", numeric.x.axis = TRUE, main = "PB CD3")
ggpar(p1, legend = "none")

  • Boxplots by testing period
ggboxplot(plot.df, x = "test", y = "pbc_cd3", fill = "relapse", main = "PB CD3")

CD15 proportions

  • Relapse vs. non-relapse
ggboxplot(plot.df, x = "relapse", y = "pbc_cd15", 
          color = "relapse", add = "jitter", main = "PB CD15") +
  stat_compare_means(label.y = 20, method = "wilcox.test") + 
  scale_y_continuous(limits = c(0,100))

  • Time plots
ggscatter(plot.df, x = "ptime", y = "pbc_cd15", col = "relapse", main = "PB CD15",
          add = "reg.line")

p1 = ggline(plot.df, x = "ptime", y = "pbc_cd15", col = "ID", 
       facet.by = "relapse", numeric.x.axis = TRUE, main = "PB CD15")
ggpar(p1, legend = "none")

  • Boxplots by testing period
ggboxplot(plot.df, x = "test", y = "pbc_cd15", fill = "relapse", main = "PB CD15")

CD34 proportions

  • Relapse vs. non-relapse
ggboxplot(plot.df, x = "relapse", y = "pbc_cd34", 
          color = "relapse", add = "jitter", main = "PB CD34") +
  stat_compare_means(label.y = 20, method = "wilcox.test") + 
  scale_y_continuous(limits = c(0,100))

  • Time plots
ggscatter(plot.df, x = "ptime", y = "pbc_cd34", col = "relapse", main = "PB CD34",
          add = "reg.line")

p1 = ggline(plot.df, x = "ptime", y = "pbc_cd34", col = "ID", 
       facet.by = "relapse", numeric.x.axis = TRUE, main = "PB CD34")
ggpar(p1, legend = "none")

  • Boxplots by testing period
ggboxplot(plot.df, x = "test", y = "pbc_cd34", fill = "relapse", main = "PB CD34")

Comparative plots

plot2.df = melt(plot.df, measure.vars = c("pbc_cdw", "pbc_cd3", "pbc_cd15", "pbc_cd34"),
                variable.name = "marker", value.name = "prop")
ggboxplot(plot2.df, x = "relapse", y = "prop", 
          fill = "marker", main = "PB All markers") +
  scale_y_continuous(limits = c(0,100))

ggboxplot(plot2.df, x = "test", y = "prop", fill = "marker", 
          facet.by = "relapse", main = "PB All markers")

  1. Stanford Medicine, ↩︎

  2. University of Utah, ↩︎