MantisBT - Rosetta
View Issue Details
0000202Rosetta[All Projects] Bad Codingpublic2013-03-12 16:562013-03-13 11:07
smlewis 
 
lowminoralways
newopen 
Trunk 
 
mm scoring, fa_standard_icoor
N/A
0000202: in fa_standard_05.2009_icoor, many atoms have X instead of VIRT for mm atom type
14:53:46 ~/rosetta/rosetta_database/chemical/residue_type_sets> diff fa_standard/residue_types/metal_ions/MG.params fa_standard_05.2009_icoor/residue_types/metal_ions/MG.params
7,9c7,9
< ATOM MG Mg2p VIRT 2.00
< ATOM V1 VIRT VIRT 0.00
< ATOM V2 VIRT VIRT 0.00
---
> ATOM MG Mg2p X 2.00
> ATOM V1 VIRT X 0.00
> ATOM V2 VIRT X 0.00


VIRT is the appropriate mm atom type for virtual atoms (and Mg, unless we get magnesium parameters) - X triggers crashes, but VIRT triggers "do not score this atom for mm purposes"

This is likely true in many places but I only checked MG.params
No tags attached.
related to 0000004resolved delucasl molfile_to_params misassigns mm atom types 
Issue History
2013-03-12 16:56smlewisNew Issue
2013-03-12 17:00delucaslNote Added: 0000163
2013-03-12 17:01delucaslRelationship addedrelated to 0000004
2013-03-13 11:05smlewisNote Added: 0000164
2013-03-13 11:07delucaslNote Edited: 0000163bug_revision_view_page.php?bugnote_id=163#r57

Notes
(0000163)
delucasl   
2013-03-12 17:00   
(edited on: 2013-03-13 11:07)
When I resolved bug 0000004 I kept the default behavior (assign mm type as X). See the close note on that report. Should we flip this switch?

(0000164)
smlewis   
2013-03-13 11:05   
These icoor files long predate bug 0000004 anyway. I say flip the switch, the MM torsion parameters treat X as bad and VIRT as nonexistent...probably nonexistence is better than badness. Generally the user is responsible for checking the atom typing anyway.

As an aside - there's no need to paste URLs - just do a # sign and then digits, and it will automagically make a link to the bug, even with strikethrough for solved bugs!