cxcalc shortestpath old and new problem

User 16a938e5e3

26-05-2009 17:11:19

Hello all,


unfortunately the problem


http://www.chemaxon.com/forum/ftopic4270.html&highlight=


still occures with the new Marvin Beans version. Moreover I get for


cxcalc --do-not-display ih shortestpath -a 4-1 coord.mol


always a "?" instead of the number of bonds. How to circumvent this?


The contents of coord.mol is:


---------------------------------------------------------------------------------------------
 OpenBabel05260918423D

 21 19  0  0  0  0  0  0  0  0999 V2000
    0.5909    0.0182    1.3594 C   0  0  0  0  0
    0.4970    1.3645    1.6290 F   0  0  0  0  0
   -0.1848   -0.4461   -0.0145 B   0  0  0  0  0
   -1.6504    0.4162   -0.0782 N   0  0  0  0  0
   -2.4199    0.4495    1.1969 C   0  0  0  0  0
    0.6144    0.0696   -1.3628 C   0  0  0  0  0
    0.1872   -0.5301   -2.5172 F   0  0  0  0  0
    0.0440   -0.6201    2.4622 F   0  0  0  0  0
    1.9062   -0.2889    1.3800 F   0  0  0  0  0
    0.4200    1.4338   -1.5682 F   0  0  0  0  0
    1.9511   -0.1177   -1.3155 F   0  0  0  0  0
   -2.5049   -0.1078   -1.1846 C   0  0  0  0  0
   -2.6992   -1.1792   -0.9835 H   0  0  0  0  0
   -3.4585    0.4600   -1.2299 H   0  0  0  0  0
   -1.9619   -0.0251   -2.1432 H   0  0  0  0  0
   -1.8390    0.9829    1.9703 H   0  0  0  0  0
   -3.3937    0.9615    1.0390 H   0  0  0  0  0
   -2.6031   -0.5924    1.5262 H   0  0  0  0  0
   -0.4672   -1.8436   -0.0555 O   0  0  0  0  0
   -0.3695   -2.2435    0.8331 H   0  0  0  0  0
   -1.3648    1.3821   -0.3202 H   0  0  0  0  0
  3  1  1  0  0  0
  1  9  1  0  0  0
  1  2  1  0  0  0
  1  8  1  0  0  0
  6  3  1  0  0  0
 19  3  1  0  0  0
 12  4  1  0  0  0
 21  4  1  0  0  0
  4  5  1  0  0  0
 17  5  1  0  0  0
  5 18  1  0  0  0
  5 16  1  0  0  0
  7  6  1  0  0  0
 10  6  1  0  0  0
  6 11  1  0  0  0
 15 12  1  0  0  0
 14 12  1  0  0  0
 12 13  1  0  0  0
 19 20  1  0  0  0
M  END


-------------------------------------------------------------------------------------------------


I appreciate this tool very much and hope there might be a solution!


 


Best regards


    rberg_01


 


 


 


 


 


 


 

ChemAxon d76e6e95eb

27-05-2009 07:12:59

I am glad that you found this tool useful. Your compound contains two separated components and the atoms you specify are located on the those separated components. So the distance of the atoms is infinity.

User 16a938e5e3

27-05-2009 09:02:31

 I see. So what is the basis of cxcalcs decision, that there are "separate" components? I suppose you agree, that there might be different concepts of "connectedness". For instance there is surley a chemical bond between the components. And then there is of course alsa certain Lewis structure without a stroke betwenn B and N, but also one with one, and so on. Is the criterion a partial charge-minimal-Lewis-structure?

ChemAxon d76e6e95eb

27-05-2009 09:49:34

Can you please attach the molecule as a file? I pasted the text in Marvin (it contains some overlapping atoms), but after cleaning it (CTRL+2) they are two separate components/fragments. No bond between the two, please see the attached images.

User 16a938e5e3

27-05-2009 10:21:44

This is as an xyz the B-N "bond" is 1.70 rather large but still a sensible range I'd say. Its probably ionic.


 

User 16a938e5e3

27-05-2009 10:26:34

And here in the corresponding .mol format (generated with babel).

ChemAxon efa1591b5a

27-05-2009 10:40:42

Hi All,


Apologies for chipping on, but it appears to me that 'you grind in two different mills' (no one in the company knew how to express this in English, this is the mirror translation of the Hungairan 'nem egy malomban őröltök').


Shortestpath is based on bear topology without any consideration of molecular geometry. In the molecule graph only covalent bonds are represented, this is the source of 'disconnectedness'. Ionic or other secondary bonds (H-bond, coordinated bonds) are not represented in the molecule graph by edges.


If you wish to calculate the distance between atoms, then you can use the geometry plugin.


Does this help or I missed something>


 


Regards,


Miklos


 

ChemAxon d76e6e95eb

27-05-2009 11:18:24

Oh, yes, you are right Miklos! Raphael, you probably need a geometrical function like distance, not a topological one, don't you? Would this solve the other problem as well?

User 16a938e5e3

27-05-2009 12:45:16

Oh sorry @all,


I missunderstood the contents of the .mol file!! It contains the connectivity information already - which in my case is produced by the "babel" which i use for the conversion of .xyz into the .mol file!! So the problem is that babel has a different view of connectivity than me.


And yes again the old problem ist the same - the connectedness is not in the .mol file!


Please excuse this stupidity!


 


So my task would now be to get a programm which can on a geometrical basis decide which atoms are bound and which not!


 


Thank you very much for patience and help!


rberg_01


 


 


 

ChemAxon d76e6e95eb

27-05-2009 17:23:06

No problem, I am glad if you got closer to the resolution.