round on survival tools

NAMESPACE

@@ -5,16 +5,21 @@ export(classification)
 export(classify)
 export(compute_likelihood)
 export(compute_posterior)
+export(cox_fit)
 export(init)
+export(kaplan_meier_fit)
 export(parameters)
 export(plot_class_fraction)
 export(plot_classification)
+export(plot_survival_analysis)
 export(posterior)
-export(survival_fit)
 import(cli)
 import(crayon)
 importFrom(dplyr,"%>%")
 importFrom(dplyr,filter)
 importFrom(dplyr,mutate)
 importFrom(dplyr,rename)
 importFrom(dplyr,select)
+importFrom(survival,Surv)
+importFrom(survival,coxph)
+importFrom(survival,survfit)

R/cox_fit.R

@@ -0,0 +1,43 @@
+# Multivariate Cox regression
+
+#' Fit multivariate Cox regression model based on INCOMMON classes.
+#'
+#' @param x A list of objects of class \code{'INCOMMON'} containing the classification results for
+#' multiple samples, as produced by using function `classify`.
+#' @param tumor_type The selected tumor type.
+#' @param gene The selected gene.
+#' @param covariates Covariates used in the multivariate regression.
+#' @return An object of class \code{coxph}.
+#' @export
+#' @importFrom dplyr filter mutate rename select %>%
+#' @importFrom survival Surv coxph
+
+cox_fit = function(x, gene, tumor_type, covariates = c('age', 'sex', 'tmb')){
+
+  x = prepare_km_fit_input(x, tumor_type, gene)
+
+  formula = 'survival::Surv(OS_MONTHS, OS_STATUS) ~ group'
+
+  for(c in covariates) {
+    what = grep(c, colnames(x), ignore.case = T, value = TRUE)
+    for(w in what) {
+      if(is.numeric(x[[w]])){
+        q =  quantile(x[w], na.rm = T)['50%']
+        x[[w]] = ifelse(x[[w]] > q, paste0('>', round(q, 0)), paste0('<=', round(q, 0)))
+        x[[w]] = factor(x[[w]])
+        x[[w]] = relevel(x[[w]], ref = grep('<=', unique(x[[w]]), value = T))
+      }
+      formula = paste(formula, w, sep = ' + ')
+    }
+  }
+
+  fit = survival::coxph(
+    formula = formula %>% as.formula(),
+    data = x %>%
+      dplyr::mutate(group = factor(group)) %>% 
+      dplyr::mutate(group = relevel(group, ref = grep('WT', unique(x$group), value = T))) %>% 
+      as.data.frame()
+  )
+
+  return(fit)
+}

R/kaplan_meier_fit.R

@@ -0,0 +1,26 @@
+#' Fit Kaplan-Meier survival model based on INCOMMON classes.
+#'
+#' @param x A list of objects of class \code{'INCOMMON'} containing the classification results for
+#' multiple samples, as produced by using function `classify`.
+#' @param tumor_type The selected tumor type.
+#' @param gene The selected gene.
+#' @return An table containing the selected gene, tumor_type, the data table used to
+#' fit the model and an object of class \code{'survfit'}.
+#' @export
+#' @importFrom dplyr filter mutate rename select %>%
+#' @importFrom survival Surv survfit
+
+kaplan_meier_fit = function(x, tumor_type, gene) {
+
+  data = prepare_km_fit_input(x, tumor_type, gene)
+  data = data %>%
+    dplyr::mutate(group = factor(group,
+                                 levels = c(
+                                   grep('WT', unique(data$group), value = T),
+                                   grep('Mutant', unique(data$group), value = T) %>% grep('with', ., invert = T, value = T),
+                                   grep('Mutant', unique(data$group), value = T) %>% grep('with', ., value = T)
+                                 )))
+  fit = survival::survfit(formula = survival::Surv(OS_MONTHS, OS_STATUS) ~ group, data = data)
+
+  return(tibble(gene = gene, tumor_type = tumor_type, data = list(data), fit = list(fit)))
+}

R/plot_survival_analysis.R

@@ -3,11 +3,12 @@
 #'
 #' @param x A list of objects of class \code{'INCOMMON'} containing the classification results for
 #' multiple samples, as produced by using function `classify`.
-#' @return An object or a list of class \code{'ggplot2'}.
+#' @return An object or a list of class \code{'ggplot2'} showing Kaplan-Meier curves and 
+#' Cox regression forest plot.
 #' @export
 #' @importFrom dplyr filter mutate rename select %>%
 
-survival_fit = function(x, tumor_type, gene, cox_covariates = c('age', 'sex', 'tmb')){
+plot_survival_analysis = function(x, tumor_type, gene, cox_covariates = c('age', 'sex', 'tmb')){
 
   km_fit = kaplan_meier_fit(x, tumor_type, gene)
 

R/utils.R

@@ -329,55 +329,6 @@ prepare_km_fit_input = function(x, tumor_type, gene){
     dplyr::select('sample', 'tumor_type', 'gene', 'gene_role', dplyr::everything())
 }
 
-# Kaplan-Meier fit
-
-kaplan_meier_fit = function(x, tumor_type, gene) {
-
-  data = prepare_km_fit_input(x, tumor_type, gene)
-  data = data %>%
-    dplyr::mutate(group = factor(group,
-                                 levels = c(
-                                   grep('WT', unique(data$group), value = T),
-                                   grep('Mutant', unique(data$group), value = T) %>% grep('with', ., invert = T, value = T),
-                                   grep('Mutant', unique(data$group), value = T) %>% grep('with', ., value = T)
-                                 )))
-  fit = survival::survfit(formula = survival::Surv(OS_MONTHS, OS_STATUS) ~ group, data = data)
-
-  return(tibble(gene = gene, tumor_type = tumor_type, data = list(data), fit = list(fit)))
-}
-
-# Multivariate Cox regression
-
-cox_fit = function(x, gene, tumor_type, covariates = c('age', 'sex', 'tmb')){
-
-  x = prepare_km_fit_input(x, tumor_type, gene)
-
-  formula = 'survival::Surv(OS_MONTHS, OS_STATUS) ~ group'
-
-  for(c in covariates) {
-    what = grep(c, colnames(x), ignore.case = T, value = TRUE)
-    for(w in what) {
-      if(is.numeric(x[[w]])){
-        q =  quantile(x[w], na.rm = T)['50%']
-        x[[w]] = ifelse(x[[w]] > q, paste0('>', round(q, 0)), paste0('<=', round(q, 0)))
-        x[[w]] = factor(x[[w]])
-        x[[w]] = relevel(x[[w]], ref = grep('<=', unique(x[[w]]), value = T))
-      }
-      formula = paste(formula, w, sep = ' + ')
-    }
-  }
-
-  fit = survival::coxph(
-    formula = formula %>% as.formula(),
-    data = x %>%
-      dplyr::mutate(group = factor(group)) %>% 
-      dplyr::mutate(group = relevel(group, ref = grep('WT', unique(x$group), value = T))) %>% 
-      as.data.frame()
-  )
-
-  return(fit)
-}
-
 
 forest_plot = function(x, tumor_types = FALSE){
 

vignettes/survival_analysis.Rmd

@@ -73,7 +73,7 @@ cox_fit(x = classified_data, tumor_type = 'PAAD', gene = 'KRAS', covariates = c(
 Plot survival fit
 
 ```{r}
-survival_fit(x = classified_data, tumor_type = 'PAAD', gene = 'KRAS', cox_covariates = c('age', 'sex', 'tmb'))
+plot_survival_analysis(x = classified_data, tumor_type = 'PAAD', gene = 'KRAS', cox_covariates = c('age', 'sex', 'tmb'))
 ```