geom_gene() displays incorrect annotations for subregions

[AdrianoWeber]

Hello,

I'm using your package to generate plots from a named list of regions. The script takes BAM files and a GTF file for a global region (downloaded from GENCODE and then subseted to my region of interest) and then plots several subregions. Here's an example of my workflow (heavily inspired from the [vignette]{https://showteeth.github.io/ggcoverage/}):

#Track loading
track.folder<-"C:/Users/adria/Documents/Unige/Masters_project/Data_PacBio/Fichiers_BAM"

#Loci definition
cyp <- "chr10:94760062-95072497"
cyp2c9 <- "chr10:94935603-94993561"
cyp2c19 <- "chr10:94760062-94858911"
cyp2c8 <- "chr10:95035045-95072497"
loci<-list(cyp=cyp, cyp2c19=cyp2c19, cyp2c9=cyp2c9, cyp2c8=cyp2c8)

# Adding genome annotations
gtf.file <- "CYP_info.gtf"
gtf.gr <- rtracklayer::import.gff(con = gtf.file, format = "gtf")

#Function
coverage.plot<-function(gene_name, threshold=10){
  #Loading datas from bam
  track.df <- LoadTrackFile(track.folder = track.folder, format = "bam",
                            region = loci[[gene_name]], extend = 2000, norm.method="None",
                            bin.size = 1)
  #Making start and end pos absolute (will be an issue for reproductibility, this step is  added because my position in BAM are relatives to the start)
  track.df$start <- track.df$start + 94760062
  track.df$end <- track.df$end + 94760062

  ##Preparing meta data
  #SampleName    Type Group
  sample.meta <- data.frame(SampleName= unique(track.df$Type))
  sample.meta$Type <- unlist(Map(function(x) strsplit(x,"\\.")[[1]][1], sample.meta$SampleName))
  sample.meta$Group <- c("POP1", "POP2","POP2","POP2","POP2","POP1","POP2","POP1","POP2","POP2","POP2","POP2","POP3", "POP2", "POP2", "POP2")

  #Changing Type
  track.df$Type<-as.factor(track.df$Type)
  levels(track.df$Type) <- sample.meta$Type
  ##Preparing mark.region for CYP genes (specific to my project)
  mark.region<-data.frame(start=c(94760063, 94773001, 94819540, 94935603, 94944159, 94970426, 95035045),
                          end=c(94773000, 94818512, 94858911, 94943679, 94970066, 94993561, 95072497),
                          label=c("CYP2C19_1", "CYP2C19_2", "CYP2C19_3", "CYP2C9_1", "CYP2C9_2", "CYP2C9_3", "CYP2C8"))
  #Doing a basic plot
  color_palette <- viridis::viridis(16)#changing color palette
  basic.coverage <- ggcoverage(data = track.df, color = color_palette, 
                               plot.type = "join", facet.key = "Type", group.key = "Type",
                               mark.region = mark.region, range.position = "out", mark.label.size=2,
                               mark.color="azure2")
  basic.coverage.thershold <- basic.coverage + geom_hline(yintercept=threshold, colour="red")#Adding the threshold used for data validation
  full.plot <- basic.coverage.thershold + geom_gene(gtf.gr=gtf.gr)
  ggsave(full.plot, file=paste0(gene_name,"_fullplot3.png"), width=20, height = 12,dpi=500)
  
}
Map(coverage.plot, names(loci))

And when generating the subplots, the coverage.plot() display the correct subregion (but not the correct X-axis) but the geom_gene() still display annotations from the start of global region so it does not correspond to what i want.
Exemple of global region plot (correct behaviour):

Exemple of working subplot (because starting at the same position than the global plot):

Exemple of problematic plot:

Is there a way for geom_gene() to adapt to the displayed region rather than defaulting to the global annotation positions and/or am I missing something trivial?
Thanks in advance for your help! 😊

Have a nice day,
Adriano