This refines the layout algorithm for annotation labels, for more relevant sorting and more precise overlap calculations. It builds on #295. Examples below highlight how this enriches the related genes kit.
Example
Previous unrefined label sorting: RAD52 was preferentially labeled, but much less well-known than the unlabeled BRCA2.
New refined label sorting: BRCA2 is much more popular than RAD52, and now labeled instead.
Also new, hovering over the full gene name shows the gene's rank, using popularity data from Gene Hints.
Another example
Excessive padding in overlap calculations hid important genes. Now, nearby genes like LDLR can be shown without colliding.
Coverage decreased (-0.5%) to 88.426% when pulling aeb966f5c6458ff4691ea1370c1dc6b5230b5908 on label-sort-overlap into 8586f29fa9531a1cd6d59322fedc5c29bfc61bb6 on master.
This refines the layout algorithm for annotation labels, for more relevant sorting and more precise overlap calculations. It builds on #295. Examples below highlight how this enriches the related genes kit.
Example
Previous unrefined label sorting: RAD52 was preferentially labeled, but much less well-known than the unlabeled BRCA2.
New refined label sorting: BRCA2 is much more popular than RAD52, and now labeled instead.
Also new, hovering over the full gene name shows the gene's rank, using popularity data from Gene Hints.
Another example
Excessive padding in overlap calculations hid important genes. Now, nearby genes like LDLR can be shown without colliding.