TIL classifier – CD8+ vs. CD4+

library(ithi.utils)
load_base_libs()

library(ithi.meta)
library(ithi.ihc)
library(ithi.xcr)

Colour palettes

pal_patient <- select_palette("patient")

Parameters

db_path <- snakemake@params$db

trb_annotated_path <- snakemake@input$trb_annotated
prevalence_option <- snakemake@params$prevalence_option
phenotype_threshold <- snakemake@params$phenotype_threshold

Metadata

db <- src_sqlite(db_path, create = FALSE)
samples <- collect(tbl(db, "samples"))
duplicates <- collect(tbl(db, "duplicates"))
xcrseq <- collect(tbl(db, "xcrseq"))

id_type <- "voa_unique"

Analysis

trb_table <- fread(trb_annotated_path)

trb_annotated <- compute_clonotype_publicity(trb_table, id_type = id_type)
trb_annotated <- annotate_phenotype(trb_annotated, phenotype_threshold)

table(trb_annotated$phenotype)

 CD4+  CD8+ 
31926 17288 

Distribution broadness

res <- wilcox.test(diversity ~ phenotype, data = subset(trb_annotated, !is.na(phenotype)))

p <- ggplot(subset(trb_annotated, !is.na(phenotype)), aes(x = phenotype, y = log10(diversity + 
    1e-05))) + stat_boxplot(geom = "errorbar", width = 0.2) + geom_boxplot(width = 0.5, 
    aes(fill = phenotype)) + theme_bw() + theme_Publication() + xlab("Phenotype") + 
    ylab("Distribution broadness") + guides(fill = FALSE)

eq <- substitute(italic(P) == p, list(p = format(res$p.value, digits = 3)))
p <- p + theme(legend.text = element_text(size = 6, family = "Helvetica"), axis.text.y = element_text(size = 6, 
    family = "Helvetica"), axis.title = element_text(size = 8, family = "Helvetica"))
p <- p + annotate("text", x = Inf, y = Inf, parse = TRUE, hjust = 1, vjust = 1, 
    label = as.character(as.expression(eq)))
print(p)

Frequency (expandedness)

res <- wilcox.test(freq ~ phenotype, data = subset(trb_annotated, !is.na(phenotype)))

p <- ggplot(subset(trb_annotated, !is.na(phenotype)), aes(x = phenotype, y = freq)) + 
    stat_boxplot(geom = "errorbar", width = 0.2) + geom_boxplot(width = 0.5, 
    aes(fill = phenotype)) + theme_bw() + theme_Publication() + xlab("Phenotype") + 
    ylab("Clonotype frequency") + guides(fill = FALSE) + scale_y_log10()

eq <- substitute(italic(P) == p, list(p = format(res$p.value, digits = 3)))
p <- p + theme(legend.text = element_text(size = 6, family = "Helvetica"), axis.text.y = element_text(size = 6, 
    family = "Helvetica"), axis.title = element_text(size = 8, family = "Helvetica"))
p <- p + annotate("text", x = Inf, y = Inf, parse = TRUE, hjust = 1, vjust = 1, 
    label = as.character(as.expression(eq)))
print(p)

So it holds up that CD8+’s are distributed more broadly (the primary finding), while the secondary finding that CD8+’s are more expanded isn’t supported by the new data.