Bugs or Limitations in the Jmol Applet
affecting FirstGlance in Jmol


These bugs/limitations are known for Jmol version 10.2.0
See also Anomalies with Selection Commands.

    FG has NO workaround for these problems.

  1. When in slab mode, Jmol reports incorrect atom identifications in hover and pickCallback. FG disables Slab in "click to select" modes (Contacts.., Hide..) and refuses to identify clicked atoms in the echo line, and disables hover when Slab is on.

  2. Wrong element. In 1H3O, THR918 lacks all atoms except .N. Jmol does not deem it to be protein (acceptable) nor hetero (correct). However, THR918.N is reported by Jmol to be oxygen by CPK color, and in the hover report.

  3. Jmol is unable to handle PDB files with two differently named residues of the same sequence number (lacking insertion codes) in the same chain.
    • 1JSA's single unnamed chain begins with MYR1 (HETATM) then GLY1 (ATOM). Jmol incorrectly deems the MYR1 to be protein, and deems the pair of residues both to have the name of the first one, "MYR". "select myr" selects the glycine atoms as well. "select myr not hetero" selects only the glycine atoms. "select myr and gly" selects 0.

  4. In rare cases, Jmol fails to connect atoms in some residues. See also anomalous atoms.
    • In 1AL4, Jmol fails to connect atoms in many sidechains including all TRP's, all VAL's, and two ALA's (all the L amino acids). However, it connects the sidechains of D amino acids DLE and DVA.
    • 2SOC is a similar case. There is one chain of 8 amino acids, 3 of which are nonstandard, and two of those are D stereoisomers. Jmol connects the atoms of the nonstandard residues, but fails to connect those of CYS, PHE, LYS, THR.
    • 1GRH has only two residues, CYS (ATOM) and HEC (HETATM). Jmol fails to connect the main chain atoms of CYS, but connects its sidechain atoms, and the atoms of HEC.

  5. Rarely, Jmol fails to recognize as protein some standard amino acids (which contain all main chain atoms). One problem that cannot be worked around is that no backbone trace is drawn for these residues. FG places spheres the same size as the backbone trace on all alpha carbons, thereby showing the "nonprotein" standard amino acids as small spheres floating in space. Also, FG has workaround code to enable it to recognize all standard amino acids as "protein" (see proteinSpt in scripts.js).
    • In 1AL4, Jmol fails to recognize as protein VAL1, ILE1^A, GLY2, and ALA3. These appear to contain all main chain atoms, but no backbone trace is drawn for these residues. Note, however, that ILE1^A is actually microheterogeneity for VAL1 (ILE1 is absent in SEQRES). In Vines, FG places spheres the same size as the backbone trace on all alpha carbons, thereby showing the "nonprotein" standard amino acids as small spheres floating in space.
    • 2SOC is a similar case. There is one chain of 8 amino acids, 3 of which are nonstandard, and two of those are D stereoisomers. Jmol fails to recognize CYS, PHE, LYS, THR as protein.
    • CYS in 1GRH (see previous item) fails to be deemed protein.

  6. Nucleic acid cartoons are drawn curiously (incorrectly?). The pentose always bonds to N1 (C, T, U) or N9 (A, G) of the base. But in the cartoon, the stem connects the backbone trace to C6 (C, T, U) or C8 (A, G) of the base. True, this carbon is closest to the phosphorus which defines the backbone.

  7. In 1VCX contains deuterated water, DOD. This is correctly displayed as "water" by Jmol. Does it contain deuterated protein? Yes, but you can't tell in Jmol because "deuterium" is undefined (also undefined in Chime). The D's are reported by Jmol as element Xx and atom 1D or 2D, and are colored as unknown. They are listed as D in the element column of the PDB file.

  8. CIF: It appears that there are some serious bugs in Jmol's interpretation of CIF files.


    FG DOES HAVE workarounds for these problems:

  9. Group names A, T, G, and C work in either case, but u works only in lower case. Workaround code in doFind() (jcontrol.js).

  10. Partially completed images appear before the script is completed, as though a "refresh" had been done midway. For example, hide all chains, and then make them reappear one at a time. This may be worse in 10.2 than it was in 10.00.48. This problem seems intermittant.

  11. Bug or feature? You can select 1, dots, then select 2, adding more dots without losing the dots for 1. But if you restrict 1, you lose the dots for 2. Also, if you "dots off" while only 1 is selected, you lose the dots for both one and two. "select none; dots off" hides all dots. Reported this one to jmol-users on August 23, 2006.

  12. In Jmol 10.2 (but not in 10.00.48) hbonds are shown by default for 1D16 (a single unnamed DNA chain with water), but not for 143D (also a single unnamed DNA chain, but hetero atoms), nor for named DNA or RNA chains such as 1D66, 104D, 1LCD, 1EVV. I did not put "hbonds off" in FG's code since the default display of hbonds seems rare, and I'd like to notice if it occurs in any other cases. OK, months later I noticed it also occurs for protein 4MDH. In FGiJ 0.995 I added an explicit "hbonds off".

Feedback to