ShixiangWang · ShixiangWang · Jun 10, 2020 · Jun 10, 2020
diff --git a/NEWS.md b/NEWS.md
@@ -1,5 +1,6 @@
 # sigminer 1.0.7
 
+- Fixed potential problem raised by unordered copy number segments ([#217](https://github.com/ShixiangWang/sigminer/issues/217)).
 - Fixed a typo, correct `MRSE` to `RMSE`.
 - Added feature in `show_sig_bootstrap_*()` for plotting aggregated values.
 - Fixed bug when use `get_groups()` for clustering.

diff --git a/R/get.R b/R/get.R
@@ -5,10 +5,25 @@
 # Get copy number list ----------------------------------------------------
 
 get_cnlist <- function(CopyNumber, ignore_chrs = NULL) {
-  if (!inherits(CopyNumber, "CopyNumber")) {
-    stop("Input must be a CopyNumber object!")
+  if (!inherits(CopyNumber, "CopyNumber") & !data.table::is.data.table(CopyNumber)) {
+    stop("Input must be a CopyNumber object or a data.table!")
+  }
+  if (is.data.frame(CopyNumber)) {
+    ## If it is a data.table
+    # order by segment start position by each chromosome in each sample
+    req_cols <- c("chromosome", "start", "end", "segVal", "sample")
+    data <- data.table::copy(CopyNumber)
+
+    if (!all(req_cols %in% colnames(data))) {
+      stop(paste0(req_cols, collapse = ","), " are necessary columns!")
+    }
+
+    data <- data[, .SD[order(.SD$start, decreasing = FALSE)], by = c("sample", "chromosome")]
+    all_cols <- colnames(data)
+    data.table::setcolorder(data, neworder = c(req_cols, setdiff(all_cols, req_cols)))
+  } else {
+    data <- data.table::copy(CopyNumber@data)
   }
-  data <- data.table::copy(CopyNumber@data)
   if (!is.null(ignore_chrs)) {
     chrs_exist <- ignore_chrs %in% unique(data$chromosome)
     if (!any(chrs_exist)) {
@@ -614,6 +629,7 @@ utils::globalVariables(
     "i",
     "chrom",
     "chromosome",
-    "segVal"
+    "segVal",
+    ".SD"
   )
 )
diff --git a/R/read_copynumber.R b/R/read_copynumber.R
@@ -347,8 +347,11 @@ read_copynumber <- function(input,
   # make sure position is numeric
   data_df$start <- as.numeric(data_df$start)
   data_df$end <- as.numeric(data_df$end)
-  # order by sample name
-  data_df <- data_df[order(data_df$sample)]
+  # order by segment start position by each chromosome in each sample
+  data_df <- data_df[, .SD[order(.SD$start, decreasing = FALSE)], by = c("sample", "chromosome")]
+  all_cols <- colnames(data_df)
+  data.table::setcolorder(data_df, neworder = c(c("chromosome", "start", "end", "segVal", "sample"),
+                                                setdiff(all_cols, c("chromosome", "start", "end", "segVal", "sample"))))
 
   if (join_adj_seg) {
     data_df <- helper_join_segments(data_df)
@@ -396,6 +399,7 @@ utils::globalVariables(
   c(
     ".",
     "N",
-    ".N"
+    ".N",
+    ".SD"
   )
 )
diff --git a/inst/test/test_joinCN.R b/inst/test/test_joinCN.R
@@ -0,0 +1,25 @@
+load(system.file("extdata", "toy_segTab.RData",
+                 package = "sigminer", mustWork = TRUE
+))
+segTabs = segTabs[rev(1:nrow(segTabs))]
+cn2 <- read_copynumber(segTabs,
+                      seg_cols = c("chromosome", "start", "end", "segVal"),
+                      genome_build = "hg19", complement = FALSE, verbose = TRUE, join_adj_seg = FALSE
+)
+
+View(cn@data)
+
+debug(sig_tally)
+cn_tally_W <- sig_tally(cn, method = "W")
+
+a <- sigminer:::get_cnlist(cn2)
+b <- sigminer:::get_cnlist(segTabs)
+
+identical(a, b)
+all.equal(a, b)
+str(segTabs$chromosome)
+
+head(a$`TCGA-99-7458-01A-11D-2035-01`)
+head(b$`TCGA-99-7458-01A-11D-2035-01`)
+
+b <- sigminer:::get_cnlist(segTabs[, -2])
diff --git a/inst/test/use_generator.R b/inst/test/use_generator.R