Dotplot view
TL;DR: The dotplot view plots alignments between two genome assemblies (or a read against a reference). Syntenic regions appear as diagonal streaks, inversions as reverse-diagonal segments, and translocations/rearrangements as off-diagonal blocks.
Opening a dotplot view
Add → Dotplot view opens the import form. It is the same one the linear synteny view uses, so see Opening a linear synteny view for its Quick start / Manual modes and the file types it accepts. Two things are dotplot-specific:
- Both axes come from the chosen track's
assemblyNames: the query horizontally and the target vertically. - Swap transposes the plot.
Cutting clutter on a busy plot
Two toolbar controls do most of the work on a whole-genome plot, and the figure below uses the second:
- The palette button's Color by menu. Query gives each sequence on the horizontal axis its own color, and Strand and Identity color by those instead. Each option carries a description of what it colors. With more than one alignment file plotted together, Distinct color per track gives each its own color, and Customize per track below it overrides the mode or pins the color for one track at a time. On a plot whose blocks are all a single pixel wide, reach for Min length first.
- Min length, in the settings menu, drops alignments shorter than the slider value. Divergent genomes align in many short fragments, and hiding them is usually what makes the syntenic blocks visible.
The settings menu, the sliders button in the header, holds the rest of what the plot looks like: Opacity and Line width for how each alignment draws, Draw CIGAR insertions/deletions for whether per-base indels are drawn inside one, Level of detail for how much of a tiered file is loaded, and Gridlines, which carries each axis' ruler ticks across the plot. At whole-genome zoom neither axis has room to number itself, so the gridlines row says none at this zoom while it stays ticked.
When a plot still reads as specks after Min length, the pair is too divergent for a whole-genome view to say anything; the next move is a smaller window.
A genome against itself
Aligning a sequence to itself puts its internal repeat structure on the plot. Tandem arrays draw as filled wedges sitting on the diagonal, and inverted repeats draw as crossings off it. Everything else on the plot is real too: the male-specific Y is dense in dispersed repeats, and at any minimum length that keeps the palindromes those blocks are drawn as well.
Both structures are described in the T2T-CHM13 Y chromosome paper (Rhie et al. 2023), which is where the coordinates the boxes use come from.
The plot says where the palindromes are. Zooming it to the boxed 4.8 Mb separates the family into four crossings, and the same track opens one of them in a linear synteny view, both panels framing one palindrome and colored by strand: the arms draw as a single minus-strand ribbon over the plus-strand match of the sequence to itself.
The gene lanes say the same thing from the annotation. Each arm carries its own copies of the same Y-linked families, RBMY1B and RBMY1A1 on one side against RBMY1D and RBMY1E on the other, CDY10P and CDY11P either side of the centre. The unpainted gaps inside the arms are where the two copies differ.
Which view to reach for is a question of how many alignments are in frame. Every ribbon spans the full height of the strip, so ribbons stack rather than separate: the same track over the whole 4.8 Mb box above is a mat, where a dotplot gives each alignment its own place on two axes. Use the plot to find the structures and a synteny view on one of them.
Producing the alignment takes one extra minimap2 flag:
minimap2 -x asm20 -P chrY.fa chrY.fa > chrY_self.paf
jbrowse make-pif chrY_self.paf --csi-P is required here: without it minimap2 keeps one primary chain per query and
marks the rest secondary, so a sequence aligned to itself returns its own
diagonal and little else. With -P every dispersed repeat is kept too, so set
Min length afterwards or the plot fills in solid at whole-chromosome zoom.
The two axes come from a synteny track's two assemblyNames, so a self
comparison needs two assembly entries pointing at the same sequence under
different names.
Navigation and interaction
Zooming: the mouse wheel always zooms both axes simultaneously. Zoom buttons in the toolbar work as well.
Panning and box-selecting: the toolbar has a mode toggle button (pan icon ↔ crosshair icon):
- Move mode (pan icon): dragging pans the view. Hold
Ctrl/Cmdwhile dragging to draw a selection box instead. - Crosshair mode (crosshair icon): dragging draws a selection box to zoom into
or open a linear synteny view. Hold
Ctrl/Cmdwhile dragging to pan instead.
Aspect ratio lock: Lock aspect ratio (same bp/px), in the ⋮ menu in the dotplot header, constrains zooming and box-selection to keep both axes at the same scale.
Opening a synteny view from a dotplot view
Click and drag to select a region, then choose Linear synteny view from the context menu to zoom into it in a new linear synteny view with both genomes as tracks.