RELEASE:
CONFIG.JS CHANGED! (FGiJ 1.98)
- now using jmolJarPathOnServer
- rcsbURLSignedHead updated
reinstate spin on in scripts.js, in makeView1Spt().
fill in the release date in top.js releasedate
fill in the release date in versions.htm
DONE delete tests from gallery
OBSOLETE: reinstate at "testing when offline" in top.js
use ?...&offline query parameter for testing files in localTestPDBFiles
DONE comment out brokenEndsReport
from end of makeMissingReport() in moltab.js (for dev/debug)
update http://proteopedia.org/wiki/index.php/Temperature_value#Missing_Residues
DELETE e.bat, j*
chmod -R o+rX *
chmod -R o+rX *
-----------------------
TESTING UNUSUAL PDB FILES
To test a custom PDB file
- prepend 'x' to filename for easy deletion later
- copy it into localPDBFiles
- mol=x???.pdb
-----------------------
JAN '14 QUICK FIXES
Under local uncertainty, mention use of Find for atoms with
specified temperature range and give example of syntax.
non-std residue help:
The "X" marks can be hidden with a checkbox in the Views or Tools tabs.
Checkbox is in all tabs! At the bottom!
Show SYNONYM field in COMPND e.g. 2soc.
-----------------------
BUGS
If XC is done first thing after load, it is blanked to black and "1D66" appears
centered at the top instead of left. Second try after a Reset works OK.
turn off hiding before unit cell or XC!
XC: don't force spin on when tool is re-clicked and we're already displaying
the XC (e.g. with spin off).
3pxg empty basket bug. Watch initial scene (huge black baskets on black
background), then change background twice (they're gone).
Phosphorus atoms in hetero seem to be marked X as non-standard nucleotides,
e.g. 4f3h. Any way to exclude these?
How much of a problem are hydrogens with large files?
TEST SOME! IF and only if a problem,
NEED
a query parameter to NOT add hydrogens for very large structures.
AND
Maybe after PDB header is processed, do a reload with pdbAddHy off?
Or fetch PDB header before full file and check the nonH atom count:
REMARK 3 NUMBER OF NON-HYDROGEN ATOMS USED IN REFINEMENT.
REMARK 3 PROTEIN ATOMS : 2080
REMARK 3 NUCLEIC ACID ATOMS : 0
REMARK 3 HETEROGEN ATOMS : 82
REMARK 3 SOLVENT ATOMS : 133
For NMR results:
Simply count the number of SEQRES lines? (13 res/line)
(Or count SEQRES lines for both XRAY and NMR to simplify)
BUT files from MakeMultimer, OPM, ??? lack SEQRES.
So maybe prefetch the entire PDB file and simply count ATOM lines?
Or can we simply look at the file size?
USE localTest PDB file with ligand 1eve for cif fetch testing.
use jmolJarPathOnServer
when localPDBCodesPathHead (config.sys) is bioinformatics.org/pdb,
then loadLigandFormat should be set to same.
Weizmann? Are ligand.cif files on oca?
amino terminal hydrogens (e.g. Leu1 in 3hyd) are not hidden in vines
with backbone trace, e.g. they are not "backbone". Probably not a bug
in Jmol.
vines does not connect Se-Se (2bc7).
unit cell script needs spin on.
better handling of alpha-carbon only e.g. 2rcj
-----------------------
MAJOR NEW FEATURES
Try:
set cartoonRibose
Should I list the new PDB validation server?
http://wwpdb.org/validation-servers.html
(you need an account submit a job -- what does it offer that is not already
available?)
Structure to full-length sequence mapping with homol mods,
intrinsic disorder, hydrophobic regions:
http://www.rcsb.org/pdb/general_information/news_publications/newsletters/2013q3/query.html#five
Use RCSB's full-length to xtl seq mapper:
http://www.rcsb.org/pdb/explore/news.do?p=general_information/news_publications/news/news_2013.html#20130827
set loadLigandFormat "URL"
"http://www.rcsb.org/pdb/files/ligand/%FILE.cif"
Ligands: open a help panel offering:
- find
- show only this ligand (sticks with thin backbone)
- show contacts to this ligand
Explore jmol "contact" for showing clashes and hbonds etc.
slab rotation? See email from Frieda, and mine to pdb-l.
CRYSTAL CONTACTS
checkbox to hide contacting
buried water using "calculate surfacedistance from {not water}"
Hydrogen bonds
Use set hbondsAngleMinimum, DistanceMaximum (item #15 at new8.htm) to show
"all" hbonds as bonds? also color hbonds energy (item #53 in new8).
In Vines, and in Contacts.
Angel's Seq3D!!!
cavities and pockets as surfaces
rebuild the atlas to use makemultimer (and temporarily elim proteopedia linx?)
???document the ability to upload a file:///path/x.pdb to signed jmol
in the URL form when you add &appl=s -- or is this a security risk?
Or how about just drag and drop a file into signed jmol, and fix FGiJ
to work properly with that? (Dropped file doesn't display when FGiJ
is running from local files. Dropping a jmol .png file just displays
the static image, replacing FGiJ.)
-------------------
LESS CLEAR BIGGIES
Need to show ALL atoms (ghostly) in the unit cell, not just the
AU. Example where their absence is problematic: 1igt.
consider set preserveState false to save memory.
also set undo false, but I don't see undo in the unsigned applet console.
add more checkboxes to S/Se to show everything mentioned
include DNA phosphate in salt bridges?
cookie to set preference to use signed applet?
Other prefs: no spin when start.
select configuration=n for altlocs, see #49 in examples-11/new.htm
1k4q has microheterogeneity (see SEQADV) and altlocs for tyr's that
are also meta-nitrotyrosine NIY. The NIY do not show up in FGiJ
1.45 as ligand (they are marked X) despite being the altloc A in
each case, 106 and 114.
Undo
Electron density maps?
Load local PDB files without sending them out? (Using signed applet)
Compare mmCIF w/ PDB and use mmCIF instead if advantages? Ask RCSB for cases.
load "..." packed; puts all atoms that occur in the unitcell. See
item #28 in new8.htm.
-----------------------
-----------
FUN
Water per residue, see xmpl in Evernote.
Can show protein mass ~protmass = {protein}.mass.sum
per chain from list of chains, and total protein and nucleic.
surfacedistance -- reporting the average, or the percentage of atoms
greater than various depths, might be interesting.
2f37 1.7 A res has occupancies of 0.5 and 0.0 with "normal" temperatures.
I noticed this because CaPTURE excludes atoms with <95% occupancy, and
then lists e.g. Arg73:A as missing sidechain atoms (not shown as S-
by FGiJ). Perhaps we need another uncertainty color scheme for
occupancy, red for 0 and blue for 1? It could be a link on the temperature
help page.
detect osbsolete entries still available, e.g. 1BEF (Murthy)
OBSLTE 08-DEC-09 1BEF
and also
EXPDTA THEORETICAL MODEL
e.g. 1dnn.
-----------
SERIOUS:
How does moltab look with a MolProbity PDB file?
OK but those files have HETATM as ATOM so lots of anomalous atoms! (water!)
Why are the protein chains in 1flo reported as DNA in Chain Details?
(It is correct in missing residues.)
table with empty-basket.png in intro needs hspace or frame (outer table?).
non-std residues should be b&s by default!
Charge should color charged hydrogens: 2kt0.
missing charges does not include C and N terminal charges.
alpha carbons are not recognized as protein (in Hide): 4ioa.
"between 1 and 16 residues are missing" not 1-16 missing. e.g 1flo
NMR: Don't offer biological unit! (or are there NMR with more than
one block in Remark 350?!) MakeMultimer does a report based on RMK 350 even
for NMR!
Add amino term charges (OXT is done).
protein and *.n and not connected(*.c)
Decrease height of something to eliminate the vertical scrollbar seen
in IE9.
Moltab fails to display when the internet is not working. Due to
visitor count?
MEDIUM:
Non-water solvent should be ligand.
sync labels show checkbox between all 4 control panels.
Give DBREF links to NCBI as well as Uniprot e.g. 2gry.
cys labels are centered.
cispeps!
Include 310 and pi helices in SS counts. However, DSSP underreports pi's
according to wikipedia.
2a65 has the record for most pi helices in one
chain but its HELIX records lack any class designations.
Jmol "calculate structure" converts them into tiny turns.
1mty has the most in a single enzyme (13).
In PDB format, in a HELIX record,
columns 39-40 distinguish 10 classes of helix by number:
CLASS NUMBER
TYPE OF HELIX (COLUMNS 39 - 40)
--------------------------------------------------------------
Right-handed alpha (default) 1
Right-handed omega 2
Right-handed pi 3
Right-handed gamma 4
Right-handed 3 - 10 5
Left-handed alpha 6
Left-handed omega 7
Left-handed gamma 8
2 - 7 ribbon/helix 9
Polyproline 10
--------------------------------------------------------------
Cols 72-76 give the length of the helix.
NOT SERIOUS:
check for CRYST1 in header and if absent, don't attempt unit cell
or crystal contacts, e.g. bilagram.pdb in Gallery.
unitcell "one moment" echo is black on a black background for reasons
unclear. See bkg.txt.
javascript alert(toggleIsDown[backgroundIndex]) is false
immediately before clicking "unit cell", yet getEchoColor() finds it
true. Yet the echo works on a white background for Crystal Contacts.
report EXPRESSION_SYSTEM_COMMON e.g. 2z7x
1dxs: missing residues -2 to 5 skipping 0 gives a spurious warning
no baskets on residue 0.
Why aren't the 2 models reported for 2vtu? (Each model has a different
altloc code A, B.)
Why does CA only e.g. 1bdx toggle with ligands+?
Document what 0.2 means for free R.
In sec struct, score % unusual helices. 1b89,1gpl have 310
helixalpha, helix310, helixpi (vs. helix)
give temp stddev!
Give av temp for each chain and for water. e.g. 3iOa (3 chains, 1 hot).
Why is the empty basket for 1b89 HUGE and BLACK?
1req: 8 non glycines lack beta carbons. Make this clickable with labels.
1bdx (CA only) when you uncheck S-, the thin backbone for protein
disappears!
When the model is alpha carbons only or P only (1bdx has both)
warn (or disallow) sidechain-requiring tools e.g. disulfides,
contacts, sb/catpi, warn in find.
Can missing residue conservation be seen at ConSurf in a sequence-display
even when there is a 3D model?
Why in 1a9x is cyg marked X a nonstandard residue (along with ADP) while orn
(hetero and protein with an alpha carbon) is not?
Also in 1a9x, ADP is marked with 2 X's per residue (2
phosphates): limit to "PA" (vs. PB)? But normal DNA/RNA is only
P. Can I select P or PA but not PB? PC in ATP?
Examples needed: long chain with multiple domains/hphob cores for polarity.
Make contacts remember backbones checkbox.
If >10 (?) S-, put orange message in introduction explaning how to turn off
and giving count. (Am I doing nucleic?)
In moltab.js makeMarkBrokenEndsSpt() generate a warning message when
neither end of a gap has coordinates.
Add broken connector and "# Missing" label to baskets.
See 2ace-missing-gaps-broken-marked folder.
>>>NEED AN UNDO<<<
TO DO:
SIGNED APPLET COOKIE AND INSTRUX FOR SAVING .JMOL FILE OR .PNG FOR APPL.
and dropping it back into the signed applet!
Surface area per residue vs. compact/pocket/choppy?
Show counts "## Missing" on broken ends.
Calculate pI, charge at pH's?
report atoms missing in NSR or ligand+ where can't show S-.
CISPEP count and click on it for clickable list.
StaRs on water etc. in 1a9x see email to jmol-users 3/30/13.
moltab sequences: mouseover a chain (FASTA, UniProt) should tooltip
the COMPND MOLECULE and SOURCE species. Example 1vu2.
make ca-only be protein, not ligand (and it can be thin backbone,change
view1 for ca-only?) e.g. 1lbg.
LOWER PRIORITY:
findIt yellow report is deselecting the form slot (so you have to click
to select for a new query).
Find does not work with e.g. c5' for deoxyribose.
Under sulfur, checkbox to spacefill nonprotein groups (within(group, ...))
containing sulfur.
Low pri bug: extra bullets for long ligands, e.g. 1kqf.
Low pri bug: 2e1c has model numbers in REMARK 465 which screw
parsing giving incorrect results in the Missing Summary table.
Also 2j9j, 1cm4.
2rgz, 2q1z crash FGiJ.
Find: a residue name containing a numeral does not work (no []).
But those are listed in the moltab and work there with [].
Example 4b14.
Low pri bug: thin backbone (all models) does not include MSE in
the backbone trace cf. 3JQO chain J, but OK in 4aek, 3uwa ???
This occurs with invert hiding. ???
======================================================================
@@ important
UNDO!!!
Use MODRES to list glycosylation sites (e.g. 1igt, 1a4g has no nonstd)
and phosphorylation sites
(e.g. 1apm). Or simply list MODRES in the nonstd/mod res help.
Use DBREF for a link to UniProt (see link in PDB file listing from OCA).
TPO, SEP are not ligand+ in 1bkx. Why not?
Make SEP, TPO, etc. show as negative charge, also OXT and amino term.
Why aren't the atoms in 1NRM connected?
thin backbone in all models, red is defined by water? eg 2ull
bug: showing all models with 2ull makes jrnl, remark1 return blank!?
3rj1 has res 4.3 with Rfree 0.37. What do I say about this?
Reliability "DUBIOUS"?
Separate S-, X, and ? checkboxes.
'First and last sequence numbers' for each chain. See my email to jmol-users
of that subject, 3/21/13.
Report % alpha and beta using SHEET, HELIX records vs. length.
======================================================================
Completion of 1.6
Revise Ppda FgiJ into Ppda page to use signed applet, write state,
comment out load command, preload same pdb file.
Disulfide bond checkboxes are not reset to default by Reset.
Make a preference (non volatile -- cookies!) to use the signed applet.
Fix zoom (using convenience button-arrows) which is exaggerated in
Jmol 12.2.30.
Bug: labeling contacts, then disulfides, messes up label size in the latter.
1f88 appears to have all occs = 1.0, but Jmol reports average 0.027
and min 0.0.
Hide: atoms with halos
Fix all references to "more views", "key resources", also in Proteopedia.
Take out "Probable" QS in Key resources since we're using Author's BU.
Fix VDW radii! spacefill 100 %Babel
set defaultvdw babel
-S, -R?
-R for residues missing in gaps or ends.
Implement "#3: Missing 7"
======================================================================
======================================================================
TOP PRIORITY FOR NEW FEATURES IN NEXT VERSION
In slab help, put a checkbox to "rotate the slab".
More views -> Preferences e.g. -S, -R, axes, unit cell, ...
Missing residue reporting code does not handle:
insertion codes in REMARK 465
model numbers in REMARK 465
3cl0 contains no REMARK 465 and S2C shows no difference between SEQRES
and ATOM sequence. Yet there are 5 gaps in the backbone, every one of which
occurs at a sequence-numbering gap. ??
306 .. 308
333 .. 335-336 .. 339-342 .. 344
433 .. 435
Hydrophobic/polar: give examples of transmembrane and link OPM.
List contacts in text.
List salt bridges and catpi in text, and checkbox to label them.
In help, explain about center of charge vs. what FG does (cf.
Proteopedia article Salt Bridges)
{*.ca}.sequence
"helixalpha", "helix310", "helixpi"
how is structure "carbohydrate" defined, vis a vis FGiJ's definition?
======================================================================
TOP PRIORITY FOR INTERIM FIXES
BUGS
Hide protein does not hide the alpha carbons in 2zuo.
Key Ext Res is blank if no PDB code e.g. uploaded homology model.
Put an expanatory message there, and/or show some resources?
Re: select protein, see jmol-users
Re: [Jmol-users] Bug: nitrogen unrecognized as protein
my reply 2/11/11
notably try
select protein,amino
instead of listing all 20 amino acids.
But is it worth making the change?
And what about nucleic?
Case with stray nitrogen in THR383: 1hjx
If you hide a chain, show ssbonds, then redisplay it, the labels are huge.
These persist even after a reset.
Why are the HETATMs in polyhedra-fujita/M24L48-crystal-simple2.pdb marked
as anomalous atoms (?)? Because it is really an XYZ file! Can I detect that
and disable anomalous atom display? (low priority!)
For alpha carbons only, use function within Jmol to detect this condition
(including *.p,*.ca), then use a javascript command to set a cao flag
in FGiJ which will change the initial scene:
labels turned off (else every CA has "S-")
Show spacefilled
colored by chain
black background
see 1sva_cao.mmol
But cf. load =1sva filter "*.ca, biomolecule 1" or somesuch?
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
checkbox to label haloed (find..) alpha carbons and phosphoruses?
"not available" from the PDB. Same message occurs if too big (out of memory).
Check whether consurf colors work on SSBONDS.
Show "chain ends" as balls using connected, cf. color monomer which restarts
blue to red for each disconnected polymer, e.g. two rainbow sequences for
2ACE. Maybe make traces translucent?
Labels for S-: omit element before residue, e.g. 1AL1.
Slab help: note that hover/click to ID may need zoom down to
work! Likely to happen only with small molecules, tho, e.g. 1al1.
Show currently loaded model in white (in troubleshooting help) to avoid
advertising the proxy.
Put 1k28.mmol on rcsb pdb directory for gallery. AND REPLACE EXAMPLE
for a URL on the main page!
In OS X FF (3.5.3/3.0.12/2.0.0.2) Jmol 11.8.14-17 does NOT have
permission to read the PDB file header from the local file.
FF DOES have permission to read this in 2.0.0.20 Win98,
and in XP 3.5.3/2.0.0.20.
First need to find out if this happens on servers!!
Apparently not since in FF w/ FGiJ 1.44 (Jmol 11.8.17)
the ConSurf example files work fine.
Also uploading a ConSurf file works fine (via OCA).
Then, if only for local files, need detection
and auto-restart with signed applet, and mention that
this problem is avoided with Safari.
All 5 tel-aviv servers have _pipe in their PDB filenames.
Why is controlsDiv undefined on reload?
The search link in Troubleshooting for HPUB entries is broken again,
goes to main page of RCSB.
Also if a non-existant code is given you go to the main page.
Re: backbone_invisible, check out 1MVM (the only one found with chain
length of 1 in RCSB). DA20:D is not ligand+. It has an unconnected o3',
but its P is connected to 3 oxygens. Now it works, and DA11:B is also
unconnected. Its "bond" is 2.98A vs. 1.57A for a real P-O bond!
Check out whether there are other .o3* not working in scripts.js.
Echo "CENTERING ON", "HIDING ON" in RED.
in doViewSpt in jcontrol cf. telAvivMode
Mention that bi.o is not always up to date and to check the beta version.
Other viewers offered by RCSB.
To work thru:
- notes.htm HI PRI
- Gallery
- check for any .htm files missed
Add writeHelpNeedingUpdate() where needed. bugsjmol.htm??
Proteopedia needs to be integrated into 1.43.
- done in "click to identify" then interpret 3-letter codes.
- done in Key External Resources.
- noted for static images in slides.htm.
vvvvvvvvvvvvvv
quaternary structure at pisa
Images and Jmol displays on the RCSB PDB website now show the complete
biological assembly for all structures--even for proteins that are
split across multiple PDB coordinate files.
1cz7 is an interesting QS case! AU has huge space between 2 proteins. RCSB
gives several BU's that look different.
For the full news item, go to:
http://pdb101.rcsb.org/pdb/static.do?p=general_information/news_publications/news/news_2009.html#20091208
^^^^^^^^^^^^^^^^
Anomalous atoms?
1b07: Gly5 (missing its O) is protein in Jmol 11.8 but not in 10.2.
1h3o: only the N of the C-terminal Thr918 is present; this remains
anomalous in Jmol 11.8.14.
1prc: there is no for0:H in the remediated file. Instead chain H begins
with FME, a NSR. Now the N of Ala333 (C-term of chain C) is its only atom,
so is anomalous.
3h1p: thr57: main chain O and CA are missing, but C, N, and
sidechain from CB are present.
/////
Non-standard residues (now shown as "X"): do all now have chains?
Currently, according to notes.htm, NSR are defined as "hetero and
(protein, nucleic)". But they could be better defined as "not
hetero and ~nonstandard", where ~nonstandard is a short list of
"(protein, nucleic) and not (ala, ..., a, ..., da, ...)".
Examples: 1a2b (GSP), 1r9s (UTP)
Ligands: APC in 1s0v is not colored as ligand in composition. This is because
ligand there is defined as not (protein, nucleic, solvent), but it is nucleic
even though HETATM. Must redefine as hetero and not solvent.
2GRY ADP is not shown as ligand, but XX!
\\\\\
could megrep for igloss and update to Proteopedia.
Mention versions 1.01-1.03 in versions.htm for >1.3.
======================================================================
Documentation completion:
Take a look at i18n for translation.
All docs with pinkbox or needingupdate.
======================================================================
Bob said Feb 2010:
Great! Eric, I hope you are integrating some of the more recent
features of Jmol, such as the ability to get away from
asynchronous message callbacks for all your queries.
jmolEvaluate() and jmolGetPropertyAsArray() should be all you
ever need. They are immediate, and they can show anything a
script can show (plus a lot more), especially with
jmolEvaluate('show("......")'). (And I have been using them for
quite some time now with total success.) Also, when you load
files, if you use
set getPdbHeader TRUE
that will be faster when you then later ask for the PDB header.
Your code should not be parsing the PDB header at all any more.
======================================================================
Why is A351 in 1BKX a "non-standard residue"?
Center atoms with halos.
Add JTAT to http://wiki.jmol.org/index.php/Recycling_Corner
Undo? Bob mentions this as an existing feature in his email of
Jan 28, 2010 "scores of new features".
NMR & XRD: When >1 model, show model count next to PDB code echo, and boldface
"All Models".
Show the name of the molecule somewhere!
In NMR All Models, do ligands same as in Proteopedia. When only 1 model,
say "Simplify" instead of "All Models" and option for color by chain.
Extract full ligand / nonstd res names from header and display somehow.
Detect alpha-carbon-only or P-only nucs and label them alpha or P instead
of S-.
See new jmol command configuration for alternate locations.
Example: arg191 in 2kin.
2A79 has S- clustered at one end. Res 2.9a, Free R 0.25.
http://www.ncbi.nlm.nih.gov/pubmed/17991435
http://www.ncbi.nlm.nih.gov/pubmed/16475981
Contacts: wireframe should be full residues, not just atoms to 7A.
Contacts needs a way to list the contacts in text. (And the Label connectors
need to be corrected for the new Jmol.)
Temperature coloring needs a color key. Also maybe a better explanation
in Ppda, with examples, re uniform B factors, e.g. 2w54, see greenblatt's
explanation. Imada-san said maybe not enough data to refine x, y, z and B factor
as well. N.B. redundancy (spots with same info) is not the same
as number of independent spots vs. number of parameters to refine,
e.g. 4 (xyz, Bfac) x number of heavy atoms.
Ramachandran (that is the command; use 'model' to switch back)
Surfaces
Load local PDB files using Pillot's method of 8/27/08:
In download help (for ConSurf) state explicitly that changes in
the scene are not saved into the PDB file, and recommend Proteopedia.
======================================================================
BUGS bugs2fix
bug after remediation: 1a2b, ribose is invisible in vines. "RNA" inhibitor?
Good example: 2vpz polysulfide reductase has both phobic and philic in same
molecule, but trp is not in interface, also consurf shows no conserved
active site.
bug in 1.39: in contacts, when residues 27-128 are selected 128
then 27 (end of chain), no * appear and contacts display is
blank. When selected 27 then 128, all is OK.
Cys (and Met?) should not be magenta in hydrophobic/polar
RCSB's sequence listing grays out residues missing in the coordinates.
Use it instead of S2C? But if you want to highlight the ends, you need
the sequence numbers -- so keep S2C?
Color key for color by uncertainty.
How will I handle MSE? 2or0.
All models appears to be using a different N->C command than N->C, cf. 2or0.
1qzv is handled well by Proteopedia (ribbons!) despite being
mostly CAO, but the mostly UNK's throw FGiJ!
2r6g has two ATP's that show as ligand in PE but as non-std
residues in FGiJ. They are properly HETATM. ??
2tun has many Ramachandran outliers (see Molprobity).
7gbp has bent trp (Rhodes, BAMBED 36:90), anomalous atoms,
non-std residue ligand AMP.
Fix the /protexpl/ links that don't use proteinexplorer.org, e.g. in notes.htm!
Detect no java with Jaim's method, and show static page w/ java.com link.
We use the function jmolCheckBrowser, part of Jmol.js
JavaScript Library.
You can implement a similar function yourself checking
the property 'navigator.javaEnabled()', as here
Checkbox to start with panels reversed, for comparator mode?
Detection of PDB files on hold. See makeBadPDBFileHelp() in
2ship/fgij/0.991and1.0
PiQSi.com (manually curated biological units) ??
Be sure to use antialiasdisplay = true for cartoons!
Issues for PDB format v3.1:
In FGiJ v 1.0 with Jmol 10.2.0:
1. a, c, g, u get RNA only.
da, dc, dt, dg get DNA only.
Change Find help.
2. Some hydrogens and oxygens in 104D that don't bother Vines
with the old format show up as floating balls in the new format.
This problem occurs in both 10.2.0 and 11.3.6.
FGiJ 1.37 for Tel-Aviv was developed with Jmol 11.1.27.
================== START CONSURF / PEPSURF:PEPITOPE / SELECTON / EPITOPIA
ConSurf reset does not turn off slab.
Pepsurf/pepitope bb remains transparent in Jmol 11.1.27
Check out dots behavior
Remove 2hu4_consurf...pdb from localpdbfiles (test in techgall).
Local consurf file problem:
If in 0tmp or slides/ummath06, using SIGNED 10.9.70, fails to recog consurf.
If in techgall, works fine.
Uploading works fine.
Either the signed applet doesn't work w/ consurf, or consurf detection
fails with a URL prefix?
If a consurf result pdb file is uploaded, strip out the old pipe block!
Currently, both come back.
Dots are hidden when halos are displayed. This is because a restrict command
is issued re: halos, after dots are refreshed. The hiding of dots is due
to a bug in Jmol, reported 8/23/06. I did not add workaround code.
Also, dots are shown in the ConSurf control panel, for hidden moieties.
Same bug affects this (can't restrict to hide without losing all dots).
There is a bug in ConSurf: insertions in seq-identical residues are not
colored, e.g. 1UCY chain N when K is processed. The insertions are hard to see
in the default view b/c gray backbones. But when you apply consurf colors
to both chains and click dotssf, they are spacefilled GRAY!
================= END CONSURF
================= BEGIN FGIJ 1.0 LIST
Jmol can be resized:
document.getElementById("jmolApplet0").width = 400
document.getElementById("jmolApplet0").height = 300
See more details in Proj Jmol -> resizing.
More Views, near Uncertainty, checkbox to show serial#, B and Occ
in click report.
add link to slides.htm (presentations) to About!
use zoomTo (fold/fxn) instead of moveto?
Add linked icons for open-access, open source: PDB, PLoS, BI.O, Creative Commons,
Connotea,
below molecular display, along with the visitor count.
Get graphics of no-java from Mz7 and add them to troubleshooting,
can't see, etc. Remember Itay Mayrose had this problem!
4HHB has tons of anomalous atoms, and two PO4's deemed nonstandard residues!
More Views: use WebMol (at RCSB) for distance matrix or interactive
ramachandran viewing, and cavities!
Water page: color water element/magenta/solvent radio buttons
(e.g. for use with charge) and then
change water graphics in help.
At the main page, pressing 'enter' for a PDB URL goes to Weizmann
inappropriately.
Reduce contrast in top part of labels_checkboxes.png
more views: checkboxes to label
radio buttons
all residues 1-letter
all AA 3-letter (nucl 1 letter)
only unusual residues (checkboxes)
-s = incomplete sidechain (B&S)
-m = incomplete mainchain (B&S)
x = non-standard residue (B&S)
? = anomalous atom (what are the consequences? it is a jmol bug)
when hydrogen is hidden explicitly in Hide.., and hiding is inverted,
all hydrogen is shown. Is this a bug?
Searching pubmed for jmol reveals some databases that should maybe be
using FGiJ?
see halos easily does not obey hiding (try it inverted!).
add cartoon to Contacts (in add'n to backbone, as req by Jeanne Hardy)
Add jeanne hardy to ack. Or, use a smoothed trace?
Use a smoothed trace for SB/CatPi too!
add a notes section on incomplete sidechains.
incomplete sidechains under more views, see incsides.spt!
In postcontacts help, change checkbox titles to span titles.
When centering, echo Atom ... centered, to avoid confusion when trying
to identify the centered atoms and forgetting to exit centering.
Make button labels bold when on (Bkg white?).
Same for current view? Well, this requires refreshing the top left panel
on every click.
TRANSLATION TO SPANISH
Check paper notes for bugs.
REMOVE PLANKS from secondary structure. Provide doc w/ snapshots showing
better renderings.
Frieda: Maybe could avoid flashing by spin off -- script -- spin on
(since no flashing when not spinning). Or, see jmol-users/hanson re new script
queue feature.
Would a less confusing hiding interface be:
First click to mark with x (click again to unmark)
Then
Hide [ ]Marked [ ]Unmarked (initially neither checked).
Clear all marks.
Exit prohibited until one checkbox checked, or marks cleared.
Checking both prohibited.
Would need complicated checking to prohibit attempts to unmark portions
of large marked entities (e.g. can't unmark residues within a marked chain).
Or, can just allow selection of residues within an already selected chain
as Contacts now does.
Jmol 10.2.0 SS bug: 1hho 52-91 is two helices, but the rocket passes thru heme
joining the ends of the two helices!
Revise seq info re: Sauder et al. emails.
Add Jmol's built-in hbonds to vines?
Would it also be possible to make the link to the PDB FASTA format sequence
A "hot link", for example:
http://www.rcsb.org/pdb/downloadFile.do?fileFormat=FASTA&compression=NO&stru
ctureId=1RTD
from Brian Foley
1H8D has lots of anomalous atoms, and one cross (TYS).
Otherwise looks like a simple X-ray protein.
Could be added to bug examples, and anomalous docs.
BUGS
Backbone only mode in Vines: re-phrase help since there is no CPK!
Bug: NSR crosses in 2GZ9 fail to hide with chain.
And they remain when hidden in contacts via bond type checkboxes.
Bug: returning to Contacts from Center fails to turn off centering!
1IO1_2.mmol (protofils), contacts to chain F (cyan), salt bridges show
only one side, nothing on opposing chain. ?
1HHO hide, center visible chains is not working! (FG 0.991)
Beginning of intro: refer to PDB button for name of molecule (if valid code!).
======================================================================
BEGIN FGIJ 2.0 LIST
======================================================================
Include nucleotide phosphates in salt bridges w/ protein cations?
Detect alpha carbon only and show backbone instead of cartoon? Cf
sv40 virus capsid in atlas.
------------------------
Royal Soc Mol Biosystems (already in adoption.htm and versions.htm)
http://www.rsc.org/Publishing/Journals/mb/News/3Dvisuals.asp
http://www.rsc.org/delivery/_ArticleLinking/DisplayHTMLArticleforfree.cfm?JournalCode=MB&Year=2006&ManuscriptID=b608703p&Iss=Advance_Article#fig2
http://www.rsc.org/Publishing/Journals/mb/News/onlinefeatures.asp
I could make upload open a new window by using the View button to
create the window containing a form, and then onLoad submit the
form (with hidden elements?).
altLoc's?
4PHV has two copies of the ligand in slightly different conformations,
with two sequence numbers (1, 2) but in the same space. Zikes!
There are no altloc codes!
Limit "quit, restart" message to Safari (?) and OSX 10.3?
see navigator.appversion in jcontrol.js
H on HOH not working correctly in contacts for 1cq2. Ah, its
deuterium! All water is deuterated, but there is H on the
non-water. Chime can't select deuterium either! Chime includes D
in "select hydrogen".
Add no pickcallback until s2j bug to bugs.
document defined terms ~xxx -- in notes?
Use .CA or .P in NSR's with a star to avoid having them be unnoticed?
Detect residues that lack sidechain atoms (e.g. Ala where larger
AA's are in SEQRES).
Show phe's near lys/arg as stick in cation pi, if even one ring atom
is within 6A. E.g. 2hhd.
----------------------------------------
Mac Safari: Clicking on the molecule while spinning (initial
view) failed to produce an atom identity report (lower left echo
line) in FG0.99/Jmol10.2 but worked in Fg0.98/Jmol10.00.46. Once
a javascript command had been sent to Jmol (e.g. click on the
Background button) identity reporting (via pickCallback) began to
work in FG0.99/Jmol10.2.
Mac Safari: THIS IS THE WELL-KNOWN BUG IN SAFARI JAVA IN OS 10.3.9.
If I ran FG0.98/Jmol10.00.46, then closed the
Jmol-containing window, then in the SAME BROWSER SESSION ran
FG0.99/Jmol10.2, no scripts sent from javascript worked. (That
is, all buttons and view-links in FG failed to have any effect.)
Also clicking on the molecule produced no atom Opening the Jmol
menu was problematic (was able to do it on my mouse's right
button but not on left button). Menu operations and commands sent
from the Jmol console worked. If I quit Safari, then ran
FG0.99/Jmol10.2 in a new Safari session, everything worked
normally, except for item #1 above.
=========================================
Seq microhet is col 17; col 27 is insertion.
1AL4 VAL1/ILE1^A is microhet w/ insertion.
=========================================
Languages! See langs.js
Adoptions page including OPM, and OCA linked to Mirrors and vv. Linked
to About.
=========================================
Image sizes:
PDB 16 x 45
OCA 17 x 46
PQS 17 x 42
=========================================
Link to ProteinGlimpse
http://homepage.mac.com/eludens/html/ProteinGlimpse.html
=========================================
Science Netwatch? (but wait till NSMB is out) netwatch@aaas.org
Application Note in Bioinformatics covering both Jmol and FGiJ?
Article in BAMBEd?
=========================================
Get PQS to adopt it? OCA?
Make a slot for the _1, _2 etc. to show PQS oligomers?
Nature's requests (see email).
Try scrollable iframe instead of div. See techinfo.htm.
To Cartoon help, add link to render non hetero as cartoon with
"color cartoon none".
Fix writePE() to use arbitrary server URL.
Advanced users: commands via console, menu.
=========================================
=========================================
Is there any need for making FGiJ downloadable? Probably not.
=========================================
tooTall help problem: The help division expands to be more than two window-
heights tall when filled with long text (Introduction).
This problem did not occur in the default no-frames fgij.htm which has
no tables (except the inner one containing the control buttons inside
the controls division).
The problem occurred when divisions were put inside a table for the
RCSB mock up.
Using the original solution (2 divs with specified % heights within
a single table cell), IE and Safari worked OK, but Gecko had the tooTall
help problem.
The following attempts at solutions failed:
f1.htm: remove 62% height in help div: induced tooTall problem in IE.
f2.htm: remove 100% height in outer div around (controls + help divs):
induced tooTall problem in IE.
f3.htm: put FGiJ in a 3-cell table within one cell of the outer RCSB table.
This worsened the tooTall problem in IE: now there are NO scrollbars in
the Intro help!
SOLUTION for f3.htm was to specify help div height in px instead of %.
(f0.htm is the original solution referred to above, retired from fgijr.htm).
----------------------------------------------------------------------
Back button on-line goes thru every operation/help (with frames only).
Need to turn off all models once displayed.
Reassurance: 8/28/05: After boot, first use on-line in IE: 95 seconds
counting, then 55 seconds with a blank black rectangle before the molecule
appeared.
Implement detection of file: vs http: and use jmol..path in jmol1.js
Report bug where incorrect strand is drawn in 1pgb-ss.pdb.
Should we enforce a square Jmol regardless of window shape?
Need to use styled div's instead of align in applet. See mail proj-jmol align.
See also jmol.js doc for CSS and mailbox CSS for how to!