UCSF Chimera Release 1.4

Release timeframe: Summer 2009

Feature Discussion

Core Features

1. Pull mmcif connectivity information from a web service. The current mmcif connectivity template file is 20MB uncompressed and constantly out of date. Rather than including the file in our release, the proposal is to include only the most common entries and pull the rest from a web service, as needed. Entries will be cached once they are downloaded, and the user will be able to download the entire file, should they desire.
2. CMS Surface code.
3. 64-bit Mac Aqua build for large volume data applications. TG
   • We had discussed having this happen when OS X 10.6 (Snow Leopard) comes out. It's supposedly slated for Q1 2009. EFP
4. Higher quality lighting and transparency using OpenGL shaders. TG
   • I guess these would be options in the Effects panel; need "effects" command for controlling all Effects (not just these new things) via script. ECM
5. Off-screen image saving (using opengl write-to-texture?) so machine does not become unusable during animation encoding. TG
6. Replace mac aqua focus-follows-mouse with click-through. Tk patch probably required. TG
7. Record usage statistics, submit with registration. TG
8. Provide our own Web services for BLAST on PDB, multiple sequence alignment (ties in with Core 4). ECM
9. Interface with comparative modeling software (ties in with Core 4). ECM
10. BILD and/or shape command ability to create ellipsoids. ECM
11. Nucleotides: saved in sessions, dialog settings maybe sticky, "nucleotides" command, possibly use in preset. ECM
12. MD Movie: checkbox in Per-Frame panel and/or Record Movie panel to recompute secondary structure (make it more obvious that ribbon display may need to be updated). ECM
13. Build Structure: allow bonding of one clump of atoms to another clump of atoms in a way that automatically repositions one clump instead of requiring the user to attempt to maneuver it into the appropriate bonding position. ECM
14. Minimize Structure: don't call Dock Prep again if all atoms already have amberType and topology has not changed. ECM
15. Read PDBPQR format and load charge and radius values as the respective attributes. ECM
    • Perhaps write PDBPQR format.
    • Perhaps read PDBQT (Autodock format) and load charge, autodockType (or something like that).
16. Make it easier to get phi,psi,omega,chi torsion values from proteins. Maybe a button on Angles/Torsions to assign the values as residue attributes? ECM
17. Reposition normal labels with mouse. EFP
18. Display multiple trajectory frames simultaneously. EFP
19. Trajectory smoothing. EFP
20. Trajectory analysis timeline plots. EFP
21. (planned/probable) Measure planes/centroids. EFP
22. (planned/probable) Rename models using Model Panel. EFP
23. (planned/probable) Coulombic surface coloring. EFP
    • Elaine would like it to generate a map.
24. (planned/probable) Finish Color Key (X3D output [needed for raytracing]; "key" option to rangecolor command). EFP

Documentation

1. Need new image-making tutorial(s) in addition to or replacing the current one. ECM

Project Related Features

Large Assemblies (Core 2)

The four major areas of core 2 are animations, molecular assembly atomic models, electron tomography maps, and single particle EM maps.

I estimate there will only be time to implement one of the following five ideas for the 1.4 release. The work would be about 1/3 time with the other 2/3 of my time devoted to collaborative, service, and dissemination activities. The visualization of membrane objects is work I am considering presenting and the electron tomography conference in Australia, Feb 1, 2009.

Molecular assemblies: Multiscale new user interface and new capabilities.

1. user-specified symmetry (e.g. C4, helix, hand-placed copies).
2. mmCIF subassembly display
3. molmap multiscale surfaces to allow quality control
4. user control over hierarchical grouping
5. reduce clutter in graphical interface (aka "button farm")
6. add multiscale command interface
7. allow writing user specified symmetry as BIOMT matrices in PDB file

Animations: Create a half dozen example movie scripts covering common cases.

1. ligand binds to molecule with morph between bound/unbound conformations
2. slice through tomography data plane by plane
3. slice through single particle map with fit pdb showing map slab and thicker pdb slab for illustrating fit
4. simultaneous morph of single particle map and pdb between 2 or more conformations
5. illustration of contact interfaces in an assembly by moving all chains radially like Mac Expose' feature
6. play a molecular dynamics trajectory (needs command interface)

Main work is to add needed commands. For instance in movie tutorial from 2008 library course 6 python scripts were required to make a simple movie: show first/last trajectory frame, calculate/play pdb morph, set subdivision quality and silhouette edges.

Tomography + single particle EM maps: Create, save, restore, use volume masks.

1. Create means make existing capabilities produce mask arrays (mask command, volume eraser, split by color) and make improved capabilities, e.g. a volume painter tool.
2. Save -- any map file format can save the masks as an array of object indices or as bit masks. Additional info to save includes object names, colors, and groupings (e.g. all microtubules). That may have to be saved in a separate file.
3. Restore may need to read from two files (mask array + object info), and need user interface to associate with primary volume data.
4. Use masks: basic capabilities to hide/show/select objects defined by mask. Ideally would like any volume tool to work also on any masked portion of a volume.

Electron tomography: Visualization of objects embedded in membranes e.g. nuclear pores in nuclear envelope, spikes in virus membranes.

1. be able to map volume gray levels onto curved membrane surfaces
2. move membrane surface along normals while viewing gray levels

3. annotate by coloring patches of surface
4. place markers on surfaces
5. extract volume regions around surface markers for subtomogram averaging (average density for several nuclear pores, virus spikes to achieve higher resolution). Requires alignment capability and handling of tens to thousands of map subregions

Single-particle EM: Improve fitting of molecules in maps. This is one of the most cited features of Chimera.

1. avoid steric clashes during fit
2. allow moving any subset of atoms, not just whole molecules
3. avoid clashes between symmetry related copies of molecules
4. allow simultaneous fit of multiple molecules

Sequence/Structure (Core 4)

• connect to our own Web service to BLAST on PDB, provide dialog for choosing which hit structures to open, whether to show pairwise sequence alignment from BLAST (in MAV), whether to superimpose hit and query structures (with Matchmaker). Input is sequence of molecule model already open in Chimera. ECM
  • (search other databases such as NR? PSI-BLAST on NR?)

• connect to our own Web service to multiple sequence alignment program (Muscle, T-Coffee, ... see ​list and links), show result in MAV. Input would be sequence alignment already open in Chimera, probably would have to reformat into unaligned multi-FASTA for server. ECM

• allow sending template structure and template/target sequence alignment to comparative modeling program, either local or Web. ECM

• given a sequence alignment and associated structure, allow showing indels on structure relative to one of the sequences in the alignment (user specifies which structure and which sequence). ECM

• allow editing sequence name. ECM

• allow specifying "sequence to add" by accession code (GI number? Uniprot accession?). Its feature annotations would automatically come along as regions. ECM
  • would we want Fetch by ID for sequences? (would we want a sequence with no structure and not aligned with any other sequences?)
  • perhaps allow specifying accession code for existing sequences, retrieve and load feature annotations as sequence regions (actually I don't know if that would work - what if the existing sequence and that database sequence were even slightly different?)

• Matchmaker option to create post-fit alignment with Match->Align and refit on that to give more meaningful # of pairs and RMSD statistics along with the better alignment. Perhaps even multiple such rounds, reporting stats and showing alignment only at the end. ECM

• MAV report all-by-all percent identities to Reply Log in form of a (text) matrix instead of a line for each pair. ECM

Most of the above aren't realistic for the 1.4 release given my work on other Chimera projects and on cluster backups. However, most of these features aren't even on my radar at this point, so it would be good to have a discussion about them in order to begin to prioritize them so that perhaps some can make the 1.5 release. EFP

Other Projects