20 November 2014 meeting, 2:00 - 3:30

Attendees

Conrad, TomG, Sam, Eric, TomF, Greg, Elaine, Scooter

Agenda

1. Updates on progress
2. Next API?
3. Final look at ​atomspec document

Next Meeting

1. Work through details of implementation plan

Action Items

1. Conrad to try using Anaconda to build chimera2
2. Greg to finish startup implementation

Results

1. Updates
   • Conrad:
     • ToolShed API: working on implementing it, but sidetracked by build problems. He finally has a useful build on Yosemite.
     • AtomSpec implementation: waiting for ToolShed completion.
   • Eric:
     • Revised UI to call new Graphics API routines
     • Waiting for ToolShed to implement command line
     • Working on translating SEQRES records into Sequences
     • Had to implement C++ logging methods to prepare for PyLogger implementation.
   • Greg:
     • Have command line working with aliases
     • Need to commit MMCIF changes
     • Looking at chimera2 startup and putting in stubs
   • Tom:
     • Working on the Graphics API implementation
     • Looked at WebGL implementation for some of Dyche Mullins data
2. Next API?
   • Logging - Eric (next meeting)
   • Triggers
   • Preferences
   • Python versions of main data objects (Molecules, etc.)
   • IO Module (File type registration)
3. AtomSpec
   • ​atomspec document
   • Remaining issues:
     • In many commands where the specification is the last argument (e.g. color), a blank specification means “all.”
       • Still true
     • Would spaces on either side of a cluster of symbols like “ &~ ” be collapsed?
       • yes
     • Asterisks (*) in input atom names will be translated to prime symbols ('), but translated back if coordinates are later saved to a file. Prime symbols in input atom names are not translated on input or output.
       • For PDB files (only), we will translate for * to ' but we will not try to translate back.
     • Any residue in PDB HETATM records (or the mmCIF equivalent) that does not already have a chain ID is assigned to chain het, unless the residue is named WAT or HOH, in which case it is assigned to chain water; residue numbers are unchanged. (See also the residue attribute isHet.) Residue specifications with chain IDs omitted will match residues in chains with single-character IDs but not residues in these special chains.
       • No. Leave it as we read it.
     • Support selection name via sel=selection_name
       • No, get rid of it
     • custom amino acid categories that have been defined with ResProp
       • No need: already supported through selection term
     • need to decide on a priority order of built-in classifications vs. target aliases vs. named selections in case of conflict, or disallow name conflicts at time of creation
       • Can't override built-in classifications
       • target aliases can either be symbolic or based on data
       • aliases can override previously defined aliases (as long as they aren't built-in)
     • Attribute names are case-sensitive; attribute values, if any, are case-sensitive if specified with ==, but not if specified with =. Attribute values containing spaces (some color names, for example) must be enclosed in double quote marks.
       • Keep single = only, but we may need to add "==" and "~" at a later time
     • Within an attribute specification, ~ can be used in place of = to indicate that the subsequent string should be treated as a regular expression.
       • See above
     • When placed before an attribute name, the caret ^{indicates that the atoms, residues, or models have not been assigned any value for the attribute. For example, :/}kdHydrophobicity designates residues (such as water or nucleic acids) that lack a Kyte-Doolittle hydrophobicity assignment.
     • OK, keep it.

Discussion

• Discussion came up about how to go about getting the first graphics up on Chimera2 screen. Specifically how much of the Hydra code to pull in.
• Will Chimera2 have triggers or not?
• Proposal for getting things built on Windows. Proposal is to use Anaconda.