search

chadadavis / sbglib

Automatically exported from code.google.com/p/sbglib
0 stars 0 forks source link

Volume-based clash #5

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago 
Subclass Domain::CofM

Original issue reported on code.google.com by chad.a.davis@gmail.com on 5 Apr 2009 at 12:54

GoogleCodeExporter commented 9 years ago 
Might be better to do this via point hashing, like Evangelia's peptide assembly.
This would allow potentially faster checks.

Original comment by chad.a.davis@gmail.com on 7 Apr 2009 at 9:46

GoogleCodeExporter commented 9 years ago 
But, a proper volume-based clash detection would make it easier to quantify the
extent of the overlap. 
Of course, that would also be possible with a point hashing approach, if the
resolution is high enough that at point occupies multiple cells and the hits are
added up. Might be overkill.

Original comment by chad.a.davis@gmail.com on 7 Apr 2009 at 9:47

GoogleCodeExporter commented 9 years ago 
And save the overlaps. 

NB this somehow broke the clash-detection logic.

Original comment by chad.a.davis@gmail.com on 7 Apr 2009 at 10:27

GoogleCodeExporter commented 9 years ago 
Support weighting, based on size of spherical representation.

Original comment by chad.a.davis@gmail.com on 7 Apr 2009 at 10:27

GoogleCodeExporter commented 9 years ago

Original comment by chad.a.davis@gmail.com on 17 Jun 2009 at 3:06

GoogleCodeExporter commented 9 years ago 
Weighting: fraction of max possible overlap, which is the volume of the smaller 
domain

Original comment by chad.a.davis@gmail.com on 23 Jun 2009 at 8:24

GoogleCodeExporter commented 9 years ago

Original comment by chad.a.davis@gmail.com on 23 Jun 2009 at 9:05

GoogleCodeExporter commented 9 years ago

Original comment by chad.a.davis@gmail.com on 20 Jul 2009 at 8:28

GoogleCodeExporter commented 9 years ago 
Also could be accomplished using the GeometricHash, which is also a more general
solution, as it's based on points, not spheres.

Original comment by chad.a.davis@gmail.com on 15 Nov 2009 at 1:04

GoogleCodeExporter commented 9 years ago 
Benchmark sphere-based clash vs Geometric Hashing of Domain CofMs

Original comment by chad.a.davis@gmail.com on 7 Dec 2009 at 1:24

GoogleCodeExporter commented 9 years ago 
Don't need to check dom being added against each existing dom.
The existing doms never move.
The complex should just (geometrically) hash existing doms. Hash the dom to
be added and look for a hit.

Original comment by chad.a.davis@gmail.com on 7 Dec 2009 at 1:25

GoogleCodeExporter commented 9 years ago

Original comment by chad.a.davis@gmail.com on 25 Oct 2010 at 11:20