Mol2db: Difference between revisions
(adding fix to multiple atom checks per level.) |
(changing to reflect new bugfix) |
||
| Line 6: | Line 6: | ||
hiergen.c uses the multiple conformations to construct the hierarchy data structure by the following algorithm: | hiergen.c uses the multiple conformations to construct the hierarchy data structure by the following algorithm: | ||
# find the | # find the largest connected set of atoms that have only one unique position, call one of those atoms the beginning of the rigid component | ||
# expand through bonds to any atoms that have the same or fewer unique positions. bonded atoms with more unique positions are recursively added as branches | # expand through bonds to any atoms that have the same or fewer unique positions. bonded atoms with more unique positions are recursively added as branches | ||
# for each branch, repeat the examining of bonded atoms and examine unique positions | # for each branch, repeat the examining of bonded atoms and examine unique positions | ||
Revision as of 21:28, 24 November 2009
mol2db is a program developed to combine multimol2 files into the Flexibase Format read by DOCK.
The purpose of this page is to document how mol2db works and any changes made to mol2db, and is not necessarily a guide to preparing files to run or running them through mol2db.
ui.c reads the inhier parameter file and sets variables
hiergen.c uses the multiple conformations to construct the hierarchy data structure by the following algorithm:
- find the largest connected set of atoms that have only one unique position, call one of those atoms the beginning of the rigid component
- expand through bonds to any atoms that have the same or fewer unique positions. bonded atoms with more unique positions are recursively added as branches
- for each branch, repeat the examining of bonded atoms and examine unique positions
this would appear to have flaws if the atom positions downstream overlap, especially since a distance tolerance is used to consider things overlapped.
confhier.c puts the conformations into the hierarchy. the potential problem here is that only one atom in each conformation in used to determine if the conformation differs from a previous one. when there are some atoms in a conformation that are the same while others are different, the results will be random based on atom ordering in the mol2 files, and can be incorrect. This has been fixed so now each conformation when added checks against all atoms in the level, not just 1.