Browse code

fixed a (serious) bug in collapseBSseq: the return object would get a methylation of 100% everywhere.

git-svn-id: file:///home/git/hedgehog.fhcrc.org/bioconductor/trunk/madman/Rpacks/bsseq@86580 bc3139a8-67e5-0310-9ffc-ced21a209358

Kasper D. Hansen authored on 19/02/2014 16:00:38
Showing3 changed files

... ...
@@ -1,5 +1,5 @@
1 1
 Package: bsseq
2
-Version: 0.11.3
2
+Version: 0.11.4
3 3
 Title: Analyze, manage and store bisulfite sequencing data
4 4
 Description: Tools for analyzing and visualizing bisulfite sequencing data
5 5
 Authors@R: c(person(c("Kasper", "Daniel"), "Hansen", role = c("aut", "cre"),
... ...
@@ -14,7 +14,7 @@ collapseBSseq <- function(BSseq, columns) {
14 14
         rowSums(getBSseq(BSseq, "M")[, ss, drop = FALSE])
15 15
     }))
16 16
     Cov <- do.call(cbind, lapply(sp, function(ss) {
17
-        rowSums(getBSseq(BSseq, "M")[, ss, drop = FALSE])
17
+        rowSums(getBSseq(BSseq, "Cov")[, ss, drop = FALSE])
18 18
     }))
19 19
     BSseq(gr = getBSseq(BSseq, "gr"), M = M, Cov = Cov, sampleNames = names(sp))
20 20
 }
... ...
@@ -11,6 +11,11 @@
11 11
     \item Fixed an issue introduced by the previous change (to
12 12
     plotRegion).  Reported (with a fix) by Tim Triche Jr
13 13
     <tim.triche@gmail.com>.
14
+    \item Fixed a serious bug in collapseBSseq reported by Julien Roux
15
+    <jroux@uchicago.edu>: it would use the Meth values instead of Cov
16
+    for the final object's Cov.  However, this will result in the return
17
+    object having a methylation of 100 percent across all loci, so
18
+    hopefully users will have seen this for themselves. 
14 19
   }
15 20
 }
16 21