Using the GRanges object for a circos plot using shiny.gosling
Setup Tracks
Track 1
We will have 3 tracks in this circos plot. The outermost layer displays the width of a sequence as a bar plot in maroon.
In the track below, we first give it an id. In the following line, we assign a title to it. This title will be displayed at the top of the rings. Next, we prepare and pass the data with track_data_gr, which takes the GRanges object, the chromosome field names and the genomic field names. We set the mark to bar as we need a simple bar plot here. We define x and y via visual_channel_x and visual_channel_y. Finally, we set the color of the bar plot and the width & height.
track1 <- add_single_track(
id = "track1",
title = "Quality",
data = track_data_gr(
gr_circos, chromosomeField = "seqnames",
genomicFields = c("start", "end")
),
mark = "bar",
x = visual_channel_x(field = "start", type = "genomic", axis = "bottom"),
y = visual_channel_y(field = "QUAL", type = "quantitative", axis = "right"),
color = visual_channel_color(
value = "#BB3114"
),
width = 700,
height = 100
)Track 2
The middle layer plots the start of the sequence to its end.
In this layer, we begin similarly by setting the id and the title of the layer. Next we pass the same data object as our data will not change. But other data can be used here. We just need to pass the appropriate chromosomeField and genomicFields. We set this as mark = "rect" as we want to fill the ring with a color if data is a certain value. These are called ideograms. We set the strokeWidth to 0.5. We set the stroke itself to have different colors based on the value in the field argument. Since “REF” is passed as field, and it has 4 values: A, C, G and T, we set the domain and range accordingly. A corresponds to color #73A97D and so on. Finally, we set the x and xe. Since this is a rect ( rectangle ) plot, it only has x and xe and no y.
track2 <- add_single_track(
id = "track2",
title = "REF",
data = track_data_gr(
gr_circos, chromosomeField = "seqnames",
genomicFields = c("start", "end")
),
mark = "rect",
strokeWidth = visual_channel_stroke_width(
value = 0.5
),
stroke = visual_channel_stroke(
field = "REF",
type = "nominal",
domain = list("A", "C", "G", "T"),
range = list("#73A97D", "#C1BB78", "#1F5E89", "#CF784B")
),
x = visual_channel_x(field = "start", type = "genomic"),
xe = visual_channel_x(field = "end", type = "genomic"),
width = 700,
height = 100
)Track 3
Finally, the innermost layer shows the changes in sequences.
For the last layer, we start off again by setting the id and the title. We set the data with track_data_gr but we now set the chromosomeField and genomicFields to to.gr columns. We also set longToWideId. This will spread the data based on event column. So the new data will have double the number of columns and will have a new column for every existing column with the suffix _2. So to.gr.start will have a corresponding column to.gr.start_2 added to the DB. This is essentially widening the dataset. We set alignment to overlay, as we want the tracks to be plotted on top of each other. We then add 2 sub-tracks to this track. The sub-tracks inherits all the data from the parent track. These tracks are set with mark = "withinLink" so that the 2 tracks are linked with each other. And we set x, xe, x1 and x1e to refer to the correct columns in the data. The first sub-track traces all variations that are not a part of ch10 and the second sub-track traces the ones that are a part of ch10. The first sub-track will trace on gray color and the second sub-track will trace in various colors based on its sequence name. In the end we also specify the width, height and the stroke width.
track3 <- add_single_track(
id = "track3",
title = "Highlight similarities of Chr 10 with others",
data = track_data_gr(
gr_circos, chromosomeField = "to.gr.seqnames",
genomicFields = c("to.gr.start", "to.gr.end"),
longToWideId = "event"
),
alignment = "overlay",
opacity = visual_channel_opacity(
value = 0.4
),
tracks = add_multi_tracks(
add_single_track(
dataTransform = track_data_transform(
type = "filter", field = "to.gr.seqnames", oneOf = list("10"), not = TRUE
),
mark = "withinLink",
x = visual_channel_x(
field = "to.gr.start", type = "genomic"
),
xe = visual_channel_x(
field = "to.gr.start_2", type = "genomic"
),
x1 = visual_channel_x(
field = "to.gr.end", type = "genomic"
),
x1e = visual_channel_x(
field = "to.gr.end_2", type = "genomic"
),
stroke = visual_channel_stroke(
value = "lightgray"
),
strokeWidth = visual_channel_stroke_width(
value = 1
)
),
add_single_track(
dataTransform = track_data_transform(
type = "filter", field = "to.gr.seqnames", oneOf = list("10")
),
mark = "withinLink",
x = visual_channel_x(
field = "to.gr.start", type = "genomic"
),
xe = visual_channel_x(
field = "to.gr.start_2", type = "genomic"
),
x1 = visual_channel_x(
field = "to.gr.end", type = "genomic"
),
x1e = visual_channel_x(
field = "to.gr.end_2", type = "genomic"
),
stroke = visual_channel_stroke(
field = "to.gr.seqnames_2",
type = "nominal",
range = c(
"#E79F00", "#029F73", "#0072B2", "#CB7AA7", "#D45E00",
"#57B4E9", "#EFE441"
)
),
strokeWidth = visual_channel_stroke_width(
value = 1.5
)
)
),
width = 700,
height = 200
)Final View
Finally we put all the 3 main tracks together along with some parameters to control the layout style, the alignment and so on.
Run App
We put all of this in a shiny app.
ui <- fluidPage(
use_gosling(),
fluidRow(
column(6, goslingOutput("gosling_plot")),
column(
1, br(), actionButton(
"download_pdf",
"PDF",
icon = icon("cloud-arrow-down")
)
)
)
)
server <- function(input, output, session) {
output$gosling_plot <- renderGosling({
gosling(
component_id = "component_1",
single_composed_track
)
})
observeEvent(input$download_pdf, {
export_pdf(component_id = "component_1")
})
}
shinyApp(ui, server)