Add insertionSummary vignette example and fix gbc factor levels

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: genecovr
 Title: Gene body coverage analysis to evaluate genome assemblies
-Version: 0.0.0.9012
+Version: 0.0.0.9013
 Authors@R: 
     person(given = "Per",
            family = "Unneberg",

R/methods-AlignmentPairsList-class.R

@@ -87,6 +87,8 @@ setMethod("geneBodyCoverage", signature = c("AlignmentPairsList"),
                                       BPPARAM=bpparam)
     }
     x <- bind_rows(lapply(gbc, data.frame), .id="id")
+    if (!is.null(names(apl)))
+        x[["id"]] <- factor(x[["id"]], levels=names(apl))
     x
 })
 

vignettes/genecovr.Rmd

@@ -134,18 +134,37 @@ plot(apl, aes(x=id, y=get_expr(enquo(cnames))), which="boxplot") + facet_wrap(.
 
 ## Plot number of indels
 
-As assembly quality improves, the number of indels go down.
+The function `insertionSummary` summarizes the number of insertions,
+either at the transcript level (default) or per alignment. The
+intuition is that as assembly quality improves, the number of indels
+go down.
+
+First we show a plot with the number of insertions per alignment. A
+consequence of this is that as a transcript may be split in multiple
+alignments, the bars are of unequal height.
 
 ```{r gbc-plot-qnuminsert}
-x <- as.data.frame(apl)
-x$cuts <- cut(x$query.NumInsert, c(-1:3, Inf), include.lowest=FALSE)
-levels(x$cuts) <- c(0:3, ">3")
-x$cuts <- factor(x$cuts, levels=c(">3", 3:0))
+x <- insertionSummary(apl, reduce=FALSE)
 ggplot(x, aes(id)) +
     geom_bar(aes(fill=cuts)) +
     scale_fill_viridis_d(name="qNumInsert", begin=1, end=0)
 ```
 
+An alternative is to summarize the number of insertions over a
+transcript. Currently, no consideration is taken to overlapping
+alignments, meaning some insertions may be counted more than once. An
+improvement would be to use the non-overlapping set of alignments with
+the fewest number of insertions.
+
+```{r gbc-plot-qnuminsert-gbc}
+x <- insertionSummary(apl)
+ggplot(x, aes(id)) +
+    geom_bar(aes(fill=cuts)) +
+    scale_fill_viridis_d(name="qNumInsert", begin=1, end=0)
+```
+
+
+
 ## Gene body coverage