The problem appears because the iodine is much larger than the H, and since you kept the same bond length, the program assumes that the iodine is at bonding distance form the nearby carbons. you should increase the I-C bond distance to fix that. I don't have experience with RasMol, but it doesn't appear to me as highly customizable. I'd suggest you to try Maestro (there's a free version that can fix those things).

I did also try increasing the bond length to that of the real molecule - same problem. If I make the bond even longer, the unreal bonds do disappear. I had not come across Maestro - I'll have to try this. Thank you.

Chemical bonds are usually NEVER specified in a PDB file.

If you look at the contents of a PDB file, you will basically see a list of all the atoms in a column, and three columns for storing the x, y, and z coordinates of each atom. The key part is that this data structure only describes a set of points in 3D space - and NOTHING about the "chemical bonds" between them. 

Whenever you load a PDB file into a molecular viewer, it is the viewing software which must make a "guess" as to which atoms are bonded to each other. Since this information is not explicitly in the PDB file, the viewer typically makes the determination based on the exact distances between each point according to some lookup table which gives the bond lengths for common bond types between common atoms. 

I don't know RasMol very well, but it may or may not be able to recognize I-C bonds based on distance. The key point is, the "truth" of your 3D file is in the PDB file - RasMol may "draw" extra bonds on the screen, but these exist nowhere outside of RasMol display, and not in the PDB file either. So the trick is to get the I atom to be the correct distance from the other C atoms as well. You can manually increase the I-"target" C atom distance to a known value from the literature, but that value may or may not correspond to the value in RasMol's own internal I-C distance table (if it exists). 

So how to get "correct" interatomic distances in your iodine containing structure? A good method, which may be what is referred to in the Maestro viewer, is applying a molecular mechanics force field, which should get you recognizable/standardized interatomic distances for the input molecular topology you specify. Alternately, the Maestro viewer may contain its own standard interatomic distance table, and move the iodine atom based on that, without calculating any physical force approximations.

Finally, what are you going to do with the PDB file next? Is it just to make a visualization? Are you going to use it in a subsequent calculation or simulation? It is possible to explicitly specific interatomic bonds using CONECT statements at the end of the PDB file; a CONECT statement listing atoms to be connected will usually "force" any molecular viewer to draw a bond between the listed atoms. 

Check out this page for more information on the PDB file format

http://deposit.rcsb.org/format-faq-v1.html

Dear  Larry,

Open your pdb file within a text editor such as gedit.

There should be in the end some "connect" information indicating which atoms are connected together. It should look like this:

CONECT   34   20                                            

CONECT   35   23   27                                       

Just try to eliminate the unnecessary connections from your file. This might already do the job.

If those annoying connections are not in the pdb file, this means that it is rasmol that perceives the bonds this way. As long as he does not write unrealistic bonds in your pdb file, the file should be usable for other simulations / calculations.

Hi Larry,

You should try minimizing this structure or at least the subset that is problematic. Many viewing softwares show the bonds based on geometric criteria. If you minimize this and still displays these bonds, then it maybe an artefact in the software. Check the generated pdb file and as long ad the CONECT card is fine, you shouldn't worry if you use it with other simulation software packages.

Also try viewing it with other viewers and see if the problem persists.

Ahmed

Thank you all for the inputs. I have indeed found that changing the interatomic distance or the atomic species can eliminate the extraneous bonds - but that is not realistic! I was not aware of the CONECT command, which is not used in any of the PDB's I have. As I am using the files for visualization only, I may be OK. l will dive in with the Swiss PDB viewer or another sophisticated tool as I need more.

wonderful responses ...

A few thoughts after seeing the different inputs...

Alexandre: Indeed, bonds are not truly indicated in the pdb file. But connection information is present, as you mention. So it seemed to me important to check whether the "undesired" bonds were present in the connect information (so in the file) or if they were just perceived by the program.

Ahmed: indeed, the structure should be energy minimized.

Larry: In the past (years ago), I experienced similar connection problems with spdbv, so I'm not sure it helps. However, there was a possibility to remove the unwanted bonds by editing the molecule. Out of curiosity: how "unrealistic" does your C-I bond length need to be in order to get rid of the unwanted bonds. C-I is normally around 2.1Å. So, when just replacing the hydrogen in the file by an Iodine, I'm not surprised that the program proposes extra bonds: those carbons are probably within that 2.1Å distance (or not too far from it...).

HI Sandro,

I have a simple PDB file that has no CONECT statements. Initially I had too short of a C-I bond length. I found that simply moving the Iodine further from the C eliminated the other bonds. However, I could not go a far as the desired 2.09Å distance, for then the Iodine was displayed with no connections. (its not quite that simple: one file it allowed that distance, in a larger PDB it did not allow that distance!) I am chalking this up to vicissitudes in the program.

Hi Larry, did you try Maestro as Eugenio proposed? What did you get out of it??? Fidele

If it's just a matter of getting a proper image, you can use the "unbond" command in PyMOL to remove the extra bonds from the display object

https://pymolwiki.org/index.php/Unbond