SMILES generation is broken in 12.0. Let's take a look at this again after I
fix that.
Note that find is not associative:
"c...@h](Cl)F".find("SMILES","CC(Cl)F")
and
"CC(Cl)F".find("SMILES","c...@h](Cl)F")
are not the same. In the first case, the result is TRUE; in the second case
it is FALSE.
Also, I have just changed the output from that function. It now returns an
array of integers indicating which atoms of the initial string were part of
the match
On Sun, Aug 8, 2010 at 11:32 PM, Otis Rothenberger <[email protected]>wrote:
> Bob,
>
> I'll send you the complete code on this if you want, but I'll simply get
> to the bottom line on a curious compare issue in this note. When I
> appended your isomer check to my existing identity check, I did the
> SMILES checks as follows:
>
> (sm2.find('SMILES',sm1)> 0)
>
> Previously, I was using a compare function that seemed strange to you in
> your earlier note to me. I think it's important to note that sm1 and sm2
> above are SMILES generated by Jmol from two models in the same frame.
>
> Since the above worked in the two model identity (and isomer) check, I
> decided to change the compare function (the strange one) that I was
> using in another routine (a model construction tutorial). In this
> routine, a single model in the window was being compared to a SMILES
> created by JME. In this routine, sm2 is the SMILES that came from Jmol
> and sm1 is the SMILES that came from JME. The routine worked for models
> with no stereochemistry. But for models with a chiral carbon, the
> routine got it backwards. So,
>
> For the strange compare function: var x =
> compare({*},{*},'SMILES',sm2,'stddev');if (x == 'NAN'), A-OK for chiral
> and non-chiral comparisons.
>
> For (sm2.find('SMILES',sm1)> 0), A-OK for non-chiral. But chiral
> backwards - i.e. R matched S and S matched R.
>
> I do understand that the SMILES/SMILES condition is a match and the
> compare statement condition is a non-match.
>
> I started using the strange compare before the SMILES to SMILES
> comparison was documented. To my way of thinking, the two approaches are
> saying the same thing. I just doubled up on the single model {*} so the
> compare function would have the correct number of arguments.
>
> In trying to track down the problem, I sent the SMILES generated by Jmol
> for (R)-2-chlorobutane to the NIH translate server. R returned S and S
> returned R
>
> So I sent (2R,3S)-2,3-dimethylpentanal. RS returned RR and RR returned RS!!
>
> I don't know how reliable NIH translate is. I was getting returned
> molfiles for input SMILES rather than returned unique SMILES because I
> can read 3D model stereochemistry much better than SMILES stereochemistry.
>
> Finally, I sent the JME SMILES for (R)-2-chlorobutane to NIH. The return
> was (R)-2-chlorobutane.
>
> On the one hand, there seems to be something wrong with Jmol generated
> SMILES, but then Jmol should get it wrong consistently.
>
> I'll triple check this stuff tomorrow, but I think the above is an
> accurate description.
>
> Otis
>
> --
> Otis Rothenberger
> chemagic.com
>
>
>
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by
>
> Make an app they can't live without
> Enter the BlackBerry Developer Challenge
> http://p.sf.net/sfu/RIM-dev2dev
> _______________________________________________
> Jmol-users mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/jmol-users
>
--
Robert M. Hanson
Professor of Chemistry
St. Olaf College
1520 St. Olaf Ave.
Northfield, MN 55057
http://www.stolaf.edu/people/hansonr
phone: 507-786-3107
If nature does not answer first what we want,
it is better to take what answer we get.
-- Josiah Willard Gibbs, Lecture XXX, Monday, February 5, 1900
------------------------------------------------------------------------------
This SF.net email is sponsored by
Make an app they can't live without
Enter the BlackBerry Developer Challenge
http://p.sf.net/sfu/RIM-dev2dev
_______________________________________________
Jmol-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/jmol-users
