Add batch id

openbiox · Dec 8, 2022 · 744880f · 744880f
1 parent 3dc4a4a
commit 744880f
Show file tree

Hide file tree

Showing 2 changed files with 22 additions and 0 deletions.
diff --git a/data-raw/update_tcga_cli.R b/data-raw/update_tcga_cli.R
@@ -0,0 +1,22 @@
+# https://zenodo.org/record/6326542#.Y4sP5ZxBzvQ
+# https://doi.org/10.1101/2021.11.01.466731
+devtools::load_all()
+
+fl = list.files("~/Downloads/TCGA_RNASeq", pattern = "rds", full.names = TRUE)
+rf = dplyr::tibble()
+
+for (i in fl) {
+  x = readRDS(i)
+  x2 = colData(x)
+  x2 = as.data.frame(x2[, c("Samples", "BatchId_mda")])
+  x2 = x2 |>
+    dplyr::as_tibble() |> 
+    dplyr::rename(sample = Samples, BatchId = BatchId_mda) |> 
+    dplyr::mutate(sample = substr(sample, 1, 15))
+  rf = dplyr::bind_rows(rf, x2)
+}
+
+tcga_clinical = tcga_clinical |> 
+  dplyr::left_join(rf, by = "sample")
+
+usethis::use_data(tcga_clinical, overwrite = TRUE)
diff --git a/data/tcga_clinical.rda b/data/tcga_clinical.rda