@@ -1,6 +1,6 @@

 Package: SplicingGraphs

 Title: Creation, manipulation, visualization of splicing graphs

-Version: 1.1.1

+Version: 1.1.2

 Author: D. Bindreither, M. Carlson, M. Morgan, H. Pages

 License: Artistic-2.0

 Description: This package allows the user to create, manipulate, and visualize

@@ -56,7 +56,6 @@ export(

     ## countReads.R:

     assignReads,

-    countReads,

     ## toy_data.R:

     toy_genes_gff,

@@ -100,7 +99,10 @@ export(

     ## rsgedgesByGene-methods.R:

     rsgedgesByTranscript,

-    rsgedgesByGene

+    rsgedgesByGene,

+

+    ## countReads.R:

+    countReads

 )

 ### Exactly the same list as above.

@@ -118,6 +120,7 @@ exportMethods(

     sgraph,

     bubbles,

     rsgedgesByTranscript,

-    rsgedgesByGene

+    rsgedgesByGene,

+    countReads

 )

@@ -136,20 +136,34 @@ assignReads <- function(sg, reads, sample.name=NA)

 ### countReads()

 ###

-### Return a DataFrame with 1 row per unique splicing graph edge and 1 column

-### per sample.

-countReads <- function(sg)

-{

-    if (!is(sg, "SplicingGraphs"))

-        stop("'sg' must be a SplicingGraphs object")

-    edges_by_gene <- sgedgesByGene(sg, with.hits.mcols=TRUE)

-    edges0 <- unlist(edges_by_gene, use.names=FALSE)

-    edges0_mcols <- mcols(edges0)

-    edges0_mcolnames <- colnames(edges0_mcols)

-    hits_idx <- grep("\\.hits$", edges0_mcolnames)

-    ans <- endoapply(edges0_mcols[hits_idx], elementLengths)

-    colnames(ans) <- sub("\\.hits$", "", colnames(ans))

-    left_cols <- edges0_mcols[ , c("sgedge_id", "ex_or_in")]

-    cbind(left_cols, ans)

-}

+setGeneric("countReads", signature="x",

+    function(x, by=c("sgedge", "rsgedge")) standardGeneric("countReads")

+)

+

+### Return a DataFrame with 1 row per splicing graph edge (or reduced

+### splicing graph edge), and 1 column per sample.

+setMethod("countReads", "SplicingGraphs",

+    function(x, by=c("sgedge", "rsgedge"))

+    {

+        by <- match.arg(by)

+        if (by == "sgedge") {

+            edges_by_gene <- sgedgesByGene(x, with.hits.mcols=TRUE)

+        } else {

+            edges_by_gene <- rsgedgesByGene(x, with.hits.mcols=TRUE)

+        }

+        edges0 <- unlist(edges_by_gene, use.names=FALSE)

+        edges0_mcols <- mcols(edges0)

+        edges0_mcolnames <- colnames(edges0_mcols)

+        hits_mcol_idx <- grep("\\.hits$", edges0_mcolnames)

+        ans <- endoapply(edges0_mcols[hits_mcol_idx], elementLengths)

+        colnames(ans) <- sub("\\.hits$", "", colnames(ans))

+        if (by == "sgedge") {

+            left_mcolnames <- c("sgedge_id", "ex_or_in")

+        } else {

+            left_mcolnames <- c("rsgedge_id", "ex_or_in")

+        }

+        left_cols <- edges0_mcols[left_mcolnames]

+        cbind(left_cols, ans)

+    }

+)

@@ -21,7 +21,6 @@

     names(txpath3) <- names(txpath)

     gene_id <- rep.int(names(txpath3), elementLengths(txpath3))

-    #tmp <- paste(gene_id, txpath3@unlistData, sep=":")

     tmp <- txpath3@unlistData

     gene_id <- gene_id[c(TRUE, FALSE)]

     from <- tmp[c(TRUE, FALSE)]

@@ -146,9 +145,19 @@

     ## Reduce hits metadata cols.

     hits_mcol_idx <- grep("hits$", colnames(edges_mcols))

     if (length(hits_mcol_idx) != 0L) {

+        ## FIXME: endoapply() on a DataFrame object is broken when applying

+        ## a function 'FUN' that modifies the nb of rows. Furthermore, the

+        ## returned object passes validitation despite being broken! Fix it

+        ## in IRanges.

         hits_mcols <- endoapply(edges_mcols[hits_mcol_idx],

                                 function(hits)

                                   unname(unique(unlistAndSplit(hits, f))))

+

+        ## Fix the broken DataFrame returned by endoapply().

+        hits_mcols@nrows <- nlevels(f)

+        hits_mcols@rownames <- NULL

+

+        ## Combine with 'ans_mcols'.

         ans_mcols <- cbind(ans_mcols, hits_mcols)

     }

@@ -192,16 +201,16 @@ setMethod("rsgedgesByGene", "SplicingGraphs",

         if (keep.dup.edges)

             stop("'keep.dup.edges=TRUE' is not supported yet, sorry")

         edges_by_gene <- sgedgesByGene(x, with.hits.mcols=with.hits.mcols)

-        all_edges <- unlist(edges_by_gene)

-        all_edges_mcols <- mcols(all_edges)

-        gene_id <- names(all_edges)

-        from <- all_edges_mcols[ , "from"]

-        to <- all_edges_mcols[ , "to"]

+        edges0 <- unlist(edges_by_gene)

+        edges0_mcols <- mcols(edges0)

+        gene_id <- names(edges0)

+        from <- edges0_mcols[ , "from"]

+        to <- edges0_mcols[ , "to"]

         ui_fqnodes <- .uninformative_fqnodes(x)

         sgedge2rsgedge_map <- .build_sgedge2rsgedge_map(gene_id, from, to,

                                                         ui_fqnodes)

         f <- factor(sgedge2rsgedge_map, levels=unique(sgedge2rsgedge_map))

-        ans_flesh <- .reduce_edges(all_edges, f)

+        ans_flesh <- .reduce_edges(edges0, f)

         ans_flesh_names <- Rle(names(ans_flesh))

         breakpoints <- cumsum(runLength(ans_flesh_names))

         ans_names <- runValue(ans_flesh_names)

@@ -133,11 +133,11 @@

         from_to_colnames <- c("end_SSid", "start_SSid")

     }

     ex_mcols <- mcols(exons)

-    ex_colnames <- colnames(ex_mcols)

-    hits_idx <- grep("hits$", ex_colnames)

-    hits_colnames <- ex_colnames[hits_idx]

-    hits_colnames <- c(from_to_colnames, hits_colnames)

-    exon_hits <- ex_mcols[ , hits_colnames, drop=FALSE]

+    ex_mcolnames <- colnames(ex_mcols)

+    hits_mcol_idx <- grep("hits$", ex_mcolnames)

+    hits_mcolnames <- ex_mcolnames[hits_mcol_idx]

+    hits_mcolnames <- c(from_to_colnames, hits_mcolnames)

+    exon_hits <- ex_mcols[ , hits_mcolnames, drop=FALSE]

     colnames(exon_hits)[1:2] <- c("from", "to")

     exon_hits

 }

@@ -160,10 +160,10 @@

     #}

     in_mcols <- mcols(introns)

     in_colnames <- colnames(in_mcols)

-    hits_idx <- grep("hits$", in_colnames)

-    hits_colnames <- in_colnames[hits_idx]

-    #hits_colnames <- c(from_to_colnames, hits_colnames)

-    intron_hits <- in_mcols[ , hits_colnames, drop=FALSE]

+    hits_mcol_idx <- grep("hits$", in_colnames)

+    hits_mcolnames <- in_colnames[hits_mcol_idx]

+    #hits_mcolnames <- c(from_to_colnames, hits_mcolnames)

+    intron_hits <- in_mcols[hits_mcolnames]

     #colnames(intron_hits)[1:2] <- c("from", "to")

     intron_hits

 }

@@ -16,7 +16,7 @@

 \usage{

 assignReads(sg, reads, sample.name=NA)

-countReads(sg)

+countReads(sg, by=c("sgedge", "rsgedge"))

 }

 \arguments{

@@ -36,6 +36,10 @@ countReads(sg)

     A single string containing the name of the sample where the reads

     are coming from.

   }

+  \item{by}{

+    Summarize by splicing graph edge (\code{"sgedge"}) or by \emph{reduced}

+    splicing graph edge (\code{"rsgedge"}).

+  }

 }

 \details{

@@ -153,11 +157,27 @@ gal

 ## junction read with 1 gap can be assigned to 2 exons and 1 intron).

 sg <- assignReads(sg, gal, sample.name="TOYREADS")

+## See the assignments to the splicing graph edges.

 edge_by_tx <- sgedgesByTranscript(sg, with.hits.mcols=TRUE)

-mcols(unlist(edge_by_tx))

+edge_data <- mcols(unlist(edge_by_tx))

+colnames(edge_data)

+head(edge_data)

+edge_data[ , c("sgedge_id", "TOYREADS.hits")]

+

+edge_by_gene <- sgedgesByGene(sg, with.hits.mcols=TRUE)

+mcols(unlist(edge_by_gene))

+

+## See the assignments to the reduced splicing graph edges.

+redge_by_gene <- rsgedgesByGene(sg, with.hits.mcols=TRUE)

+mcols(unlist(redge_by_gene))

 ## ---------------------------------------------------------------------

 ## 4. Count the number of reads per splicing graph edge

 ## ---------------------------------------------------------------------

 countReads(sg)

+

+## ---------------------------------------------------------------------

+## 5. Count the number of reads per reduced splicing graph edge

+## ---------------------------------------------------------------------

+countReads(sg, by="rsgedge")

 }