Selection Issues in Jmol Command Scripts
with Proposals for Improvements to Jmol

Problems selecting ligands+, non-standard residues, and protein are indicated in red.
Results in black are OK.
Results below are for Jmol 10.2.0.



 
PDB Code
Group Name or Atom
=(protein,nucleic)
Chain
Hetero
Ligand
Displayed as
Ligands+ in FGiJ
Displayed as
NSR* in FGiJ
2BBN
(no NSR*)
Ca
YES
A
Yes
Yes
Yes
No
1TCO
Ca
FE, ZN, PO4, MYR, FK5: OK
YES
(none)
Yes
Yes
Yes
No
1EVV
NSR*
Mg, SPM
Yes
No
A
(none)
Yes
Yes
Yes
Yes
No
Yes
Yes
No
1H3O
THR918.N (lacks all other atoms)
MSE
No
Yes
A
A
No
Yes
No
Yes
Yes
No
No
Yes
1LCD
(no NSR*)
Na
No
(none)
Yes
Yes
Yes
No
1K28
K
PO4
MSE
No
No
Yes
(none)
(none)
Yes
Yes
Yes
Yes
Yes
No**
Yes
Yes
Yes
No
No
No
Yes
310D
Mg
CH3
No
No
(none)
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
1B07
GLY5 (notes 1,3) lacks =O on .C
PYL
NO
No
C
C
No
Yes
No
Yes
No
Yes
No
No
1AL4
VAL1,ILE1^A,GLY2,ALA3 (note 3)
DLE, DVA  (see bugsjmol.htm)
FOR
POL
NO
Yes
No
No
A,B
A,B
A,B
(none)
No
Yes
Yes
Yes
No
Yes
Yes
Yes
No
No
Yes
Yes
No
Yes
No
No
1UC5
MET79:M
K, NH4, PGR, CNC: OK
NO
M
No
No
No
No
1JSA
MYR1 (note 2)
YES
(none)
No
No
NO
YES
2SOC
CYS, PHE, LYS, THR (note 3)
DPN, DTR, THO
NO
Yes
(none)
(none)
No
Yes
No
Yes
No
No
No
Yes
1GRH
CYS (note 3)
HEC
NO
No
(none)
(none)
No
Yes
No
Yes
No
Yes
No
No
1PRC
ALA333.N:C (lacks all other atoms)
FOR0:H ("ATOM"! error in PDB)
see 1PRC below
No
No
C
H
No
No
No
No
No
No
No
No

* NSR = Non-Standard Residues (amino acids or nucleotides)
**Solvent includes PO4 in 1K28.

Note 1: Standard residues that are in ATOM records, yet not deemed protein by Jmol sometimes lack atoms:

Note 2: 1JSA's single unnamed chain begins with MYR1 (HETATM) then GLY1 (ATOM). Jmol incorrectly deems the MYR1 to be protein. "select myr" selects the glycine atoms as well. "select myr not hetero" selects only the glycine atoms. "select myr and gly" selects 0.

Note 3: FG shows these with dots as "anomalous atoms". See notes.htm#anomalous

Other test PDB's:

Preamble to proposals for changes in Jmol

RasMol and Chime divide all atoms into ATOM (protein,nucleic) or HETATM (hetero). Hetero is further sub-divided into solvent (water/HOH, DOD, and inorganic sulfate and phosphate) with all non-solvent being subsumed in the term "ligand". No HETATM can be protein or nucleic. All atoms are either (protein,nucleic) or hetero; no atoms are neither.

In Jmol, (protein,nucleic) is allowed to include HETATM's that meet certain criteria. This enables hetero non-standard residues (NSR) to be included in backbone traces, and excluded from "ligand", which is desirable. It also allows the selection of NSR with "(hetero and (protein,nucleic))" [when the authors of the PDB file have designated the entire NSR as HETATM, as in 1BKX, but not when the main-chain portions are designated ATOM, as in 1APM]. Being able to distinguish NSR from ligand is desirable.

In some PDB files, Jmol deems some "orphan" atoms neither (protein,nucleic) nor hetero, notably standard amino acids that lack the usual atoms (1TCO). This seems acceptable, since such atoms can be detected by selecting "not (protein,nucleic,hetero)".

See further discussion in notes.htm#nsr

Proposals for changes in Jmol:

  1. If something is (protein,nucleic), it cannot also be ligand. (As stated in the preamble above, in RasMol and Chime, these are mutually exclusive, which seems a good thing.) This change removes NSR from ligand, enabling us to exclude ligand from the NSR display. This change seems desirable, though there is no example above where this is crucial (but only because I have not used the term ligand in selections in FGiJ). In FG, NSR are selected with "(protein,nucleic) and hetero"; ligand with "not (protein,nucleic,water)".
  2. Any atom that belongs to an ATOM (not HETATM) group with one of the standard amino acid or nucleotide names should be (protein,nucleic), unless it lacks main chain atoms. (It would be PDB-format-illegal for atoms to have a standard residue name and be HETATM.)  This change corrects the present failure to deem standard residues as protein in 1B07 and 1UC5. [The reason why VAL1 in 1AL4, and 4 standard amino acids in 2SOC, and CYS in 1GRH are deemed nonprotein is unclear to me. None have missing main-chain atoms.] As stated elsewhere in this document, the fact that THR918.N in 1H3O lacks all atoms except .N is reason enough for Jmol not to deem these .N's "protein".
  3. It may be necessary to permit two distinct residues in the same chain to have the same number, yet different characteristics. One example is MYR1 and GLY1 in 1JSA (see Note 2 above). Another is 1DPO, which has two different amino acids at positions 184, 188, and 221 without insertion codes. This is probably very rare, and might be illegal in PDB format?
  4. 918.N in 1H3O is reported to be oxygen by CPK color, and in the hover report (not shown in table above). This bug needs to be corrected.
It would be more appropriate to display the CH3 in 310D and the PYL in 1B07 as NSR (rather than ligand), but this is probably difficult to achieve, and we can live with the present situation.