[Biopython] arrowhead width

[Peter]

On Mon, Jun 8, 2009 at 11:32 PM, Iddo Friedberg<idoerg at gmail.com> wrote:
> Hopefully the attached png clarifies things.

Much clearer :)

> The arrow shaft can be narrowed using its own argument as you pointed out. I
> would like to make the arrowhead width narrower, the part perpendicular to
> the track . But you can only be defined using the box rather than an
> argument such as arrowhead_width?

Right now you can't do what you want to an individual feature.
However, you can do it to *all* the features on the track, by reducing
the height of the track itself. Do you have something specific in
mind, or just a desire to tweak the image? I suppose it could be
useful, and the code wouldn't be too bad.

Changing the height of the bounding box has implications on its
vertical (here radial) position.  Something I have discussed with
Leighton is allowing the height of a feature to be set (defaulting to
1.0, meaning the full vertical space of the track as now). This would
change the height of the BOX sigil, or the height of the bounding box
for the ARROW sigil - indirectly doing what you want but also "moving"
the arrow closer to the center of the track. I have found this allows
some interesting ways to represent microarray expression (using a BOX
sigil looks better than the arrows), but this kind of change is best
considered with a long term plan in mind...

In the long term, some way to have multiple feature at different
vertical offsets may be needed (perhaps with different vertical
heights) - but this is quite a big change. e.g. Showing CDS features
with their exons at different vertical heights for different frames
would be nice. Also, automatically laying out a diagram "bumping"
features to avoid visual overlap. These variants might all be regarded
as "sub feature tracks". However, at the moment I have other
priorities.

Peter

P.S. Circular diagrams look better with some "dead space" in the
center (as done in the tutorial by effectively having some empty
tracks). I've wondered about having an extra option for a "dead space
radius", this seems cleaner!