conjugated bonds alignment

Hi there,





when I tried to construct a benzoate structure (yes - to construct a ligand ;) ), there are two ways, how MarvinSketch will produce the output.





One, where the dashed lines next to the continuous are on the same site and therefore connected (nice) and one, where they lie on different sites (not so nice).





Unfortunately, I just had the "nice" case only once and arrive at the not so nice case usually.








I catched the string of the two examples - may be someone can help me to understand, why one is this and the other is another way?





Thanks a lot,





pheidrias

Code:

<?xml version="1.0" ?> <cml> <MDocument>   <MChemicalStruct>     <molecule molID="m1">       <atomArray           atomID="a1 a2 a3 a4 a5 a6 a7 a8 a9 a10 a11 a12 a13 a14 a15 a16 a17 a18"           elementType="C C C C C C C O O C C C C C C C O O"           x2="-13.066066983832455 -13.066066983832455 -14.39974610566049 -15.733425227488526 -15.733425227488528 -14.399746105660492 -14.39974610566049 -15.733425227488524 -13.066066983832457 -21.669998168945312 -23.00366750999357 -23.00366750999357 -21.669998168945312 -20.336328827897056 -20.336328827897056 -21.669998168945312 -20.33631904711728 -23.00367729077335"           y2="-0.1848107026880914 -1.7248107026880959 -2.4948107026880955 -1.7248107026880959 -0.18481070268809496 0.5851892973119064 -4.034810702688096 -4.804810702688096 -4.804810702688096 0.6600340050957421 -0.10998293546329663 -1.650016816581381 -2.42003375714042 -1.650016816581381 -0.10998293546329663 -3.96003375714042 -4.730033757140421 -4.730033757140418"           />       <bondArray>         <bond atomRefs2="a1 a2" order="1" />         <bond atomRefs2="a1 a6" order="2" />         <bond atomRefs2="a2 a3" order="2" />         <bond atomRefs2="a3 a4" order="1" />         <bond atomRefs2="a3 a7" order="1" />         <bond atomRefs2="a4 a5" order="2" />         <bond atomRefs2="a5 a6" order="1" />         <bond atomRefs2="a8 a7" order="1" queryType="SA" />         <bond atomRefs2="a7 a9" order="1" queryType="SA" />         <bond atomRefs2="a10 a11" order="1" />         <bond atomRefs2="a10 a15" order="2" />         <bond atomRefs2="a11 a12" order="2" />         <bond atomRefs2="a12 a13" order="1" />         <bond atomRefs2="a13 a14" order="2" />         <bond atomRefs2="a14 a15" order="1" />         <bond atomRefs2="a13 a16" order="1" />         <bond atomRefs2="a16 a17" order="1" queryType="SA" />         <bond atomRefs2="a16 a18" order="1" queryType="SA" />       </bondArray>     </molecule>   </MChemicalStruct> </MDocument> </cml>

Isn't anybody out there having similar problems?





The cml-string should be reproducable - I just tried it myself again...








Or is there something I've been overlooking?














Thanks a lot for some help,





pheidrias

Hi Guntram,





Sorry for the delay in our answer. The MRV you sent can be load, no problem, and I see the difference, but I could not reproduce the drawing of the "nice" structure for long, because - as you also mentioned - with normal drawing techniques the result is always the "not so nice" one.





Now I found a way that will result in the "nice" structure, and although I know it is not the proper solution of the problem, I share it with you:

• Select the 'Single or Aromatic' bond tool
• Draw a bond
• Draw another bon separated from the first one
• Select one of the bonds, and rotate to the requested "nice" position
• Select one of the bonds and move towards the other bond until one of the carbon atoms of it is merged to a carbon atom of the other
• now you can continue the drawing with normally (e.g. by attaching the phenyl ring)

I know that the proper solution would be to be able to change the position of the dashed line from one side of the bond to the other e.g. by consecutive clicks with the 'Single or Aromatic' bond tool, so I create a feature request task for that, but to be honest, I don't know when we will have capacity to implement it.





Best regards,





Akos

Hi Akos,





thanks for the reply and the workaround. This one works and the "concecutive click" is surely something good to have on the "feature request list" (is this list open to everybody?).





MarvinSketch is getting very close to a leading position for the everyday drawing of molecules! I will take care myself to spread it over the world :-).














Thanks,





pheidrias

Hi Guntram,





It's a good news that you like MarvinSketch more and more :) We do our best to continuously improve it.





No, the feature request list is not public, but if you have a feature request, I tell you whether it is on the list or not ;-)





Best regards,





Akos

Hi Akos,





there is one main thing for the feature request/repair list: The 2D-Clean-Function!





It always changes the whole orientation of the molecule!





Although it doesn't keep stereo-implications properly. E.g, in vinylbenzole it changes the orientation of the double-bond.





Is it on the list? Would be a big plus in fighting ChemDraw ;-).

Hi,





The 2D clean must result in the same coordinates from the same structural info, this is why it sometimes changes the orientation. I suggest to select only the part of the molecule you would like to clean, then it the orientation of the whole molecule will not be effected.





Best regards,





Akos

Hi Akos!





But partial optimization still has the problem of changing the stereochemistry!














See this example of salicylaldehyde, where this structure will be "optimised" to another stereochemistry - which does matter, if you know, that C=O and OH will give a hydrogen-bond...





In this moment I see, that the pasted structure behaves correctly whereas the copied structure in the same document switches orientation of the aldehyde-function to head away from the the OH-substituted side of the phenylring...

Code:

<?xml version="1.0" ?> <cml> <MDocument>   <MChemicalStruct>     <molecule molID="m1">       <atomArray           atomID="a1 a2 a3 a4 a5 a6 a7 a8 a9"           elementType="C C C C C C C O O"           x2="-15.943338365905165 -17.2770174877332 -17.277017487733204 -15.943338365905168 -14.609659244077132 -14.609659244077129 -15.943338365905161 -14.362016572205853 -13.275980122249091"           y2="-30.005556490156415 -30.77555649015642 -32.31555649015642 -33.08555649015642 -32.315556490156425 -30.775556490156426 -28.465556490156416 -27.69555649015642 -30.005556490156422"           />       <bondArray>         <bond atomRefs2="a1 a2" order="1" />         <bond atomRefs2="a2 a3" order="2" />         <bond atomRefs2="a3 a4" order="1" />         <bond atomRefs2="a4 a5" order="2" />         <bond atomRefs2="a5 a6" order="1" />         <bond atomRefs2="a1 a6" order="2" />         <bond atomRefs2="a7 a1" order="1" />         <bond atomRefs2="a8 a7" order="2" />         <bond atomRefs2="a6 a9" order="1" />       </bondArray>     </molecule>   </MChemicalStruct> </MDocument> </cml>

Hi,





I don't think that hydrogen-bonding is taken into account during the 2D clean (although if I clean the example molecule you sent the carbonyl group turns toward the hydroxyl group, which is correct), but I assign this topic to the expert of 2D clean algorithm to give you the final answer.





Best regards,





Akos

Hi





Hydrogen bonding is not taken into account during the 2D clean.

Quote:

See this example of salicylaldehyde, where this structure will be "optimised" to another stereochemistry

The aldehyde group can freely rotate around the single bond so I don't think it is a bug in the clean2D.














Andras

Hello together!





I was afraid of these answers.





I only mentioned hydrogen bonding for reasons of chemical understanding!





For drawing reasons, the orientation of the aldehyde group either on the side of the hydroxy group or on the other side is an important thing!





Another example (without hydrogen bonding or any other (weak) chemical interactions) would be some extended conjugated system (two ore more pi bonds in conjunction), where there is no more free rotation around the "single" bond - because this bond (drawn as a single bond) is involved in the pi-bonding.














Even without this implication of chemical sense or not, the described behaviour is simply a 3D-rotation/flip of the molecule (e.g. 180° around the x-axis) in a 2D-drawing environment, which is obviously forbidden ;-).














So please do take care of this problem - a proper clean-up function is a must-feature (and the only reason, why I'm using ChemDraw!), thanks,





pheidrias

Hi,








I agree that extended conjugated system has high energy barrier around the "single" bond rotation (as they are not single bonds any more but conjugatet ones).





So let me explain the salicylaldehyde example.





Clean function works from the following information if given:





- atom types





- bond types





- connection information





- double bond stereo





- chirality info








(As you can see extended conjugated information is not used by clean2D.)





So saliciylaldehyde has the following information:





benzene ring + aldehyde group + hydroxi group in ortho position





from this information we have two possibilities for the aldehyde group





clean chooses one.





Consider the SMILES of the molecule: OC1=C(C=O)C=CC=C1





The aldehyde group orientation is not coded. Actually clean 2D function was originally designed to depict SMILES strings. So, if I understand well you suggest to determine information from extended conjugated system (if possible if the molecule is given in 2D or 3D) and use them as well. But if the molecule is given in 2D or 3D why would someone use clean function?








One more question in my mind. If I understand well you say that the energy barrier for the rotation of aldehyde group is quite high. I have no data about that but for the first sight I think it can easily change from conformation to the other in solution under room temperature. Can you give me some reference about this energy barrier?














Andras

Hi Andras,





a quickly done ab initio calculation DFT/RB3LYP/6-31-G(d) (solvent water) says the following:





The conformation (C) with the carbonyl-oxygen on the hydroxy-side and with the possibilty to make a hydrogen bond is the most stable conformation (= 0 kcal/mole).





The other planar conformation (T) is about 6 kcal/mole less stable  than (C).








The conformation with the formylgroup rotated 90 degrees is about 17 kcal/mole less stable than conformation (C). (This 90 degree conformation is not optimized).





Regards, Hans-Ulrich

Dear Hans-Ulrich,





Thank you for the calculation results.





It is said that usually two species cannot be separated in room temperature if the energy barrier is below 100 kJ /mol. In this case 17 kcal/mol (which is 71 kJ/mol) is below this limit so my opinion is that the formyl group can rotate around the single bond.





Dear pheidrias please let me know your comments about this.





Andras

Quote:

But if the molecule is given in 2D or 3D why would someone use clean function?

Hi Andras,





simply said: The clean-up function will be needed to have a "nice" output. Maybe one would refer to this as a "beautify"-function then.





See the attached molecule, which came out like this from click-by-click drawing.





It's easy to be seen, that the "upper" phenyl-ring is not in a parallel alignment to the canvas (or the other phenylring), which is of course very desireable if you don't want to have those ugly stair-effects when taking a closer look to the pixels...





So how to fix it?


Selecting the whole molecule will flip it --> but maybe this alignment (Hydroxy to the right) is important to hightlight a special feature.





Selecting only the phenyl ring and "Clean Up" will result something I won't talk about (look for yourself).





Selecting phenyl ring including the nitrogen atom will the other stereochemistry.





Selecting Phenyring + Imine (C=N) group does likewise.

Code:

<?xml version="1.0" ?> <cml> <MDocument>   <MChemicalStruct>     <molecule molID="m1">       <atomArray           atomID="a1 a2 a3 a4 a5 a6 a7 a8 a9 a10 a11 a12 a13 a14 a15"           elementType="C C C C C C C N O C C C C C C"           x2="-6.820000171661377 -8.153669512709635 -8.153669512709635 -6.820000171661377 -5.486330830613118 -5.486330830613118 -6.820000171661377 -5.486321049833341 -4.152646818475888 -5.486321049833341 -4.090615324890851 -3.9563925281254453 -5.217875456302531 -6.613581181245021 -6.747803978010426"           y2="3.850033823897622 3.0800168833385815 1.5399830022205006 0.7699660616614601 1.5399830022205006 3.0800168833385815 5.390033823897622 6.160033823897622 3.850008413012473 7.7000338238976225 8.350883675524523 9.88505726277601 10.768380998400593 10.117531146773691 8.583357559522208"           />       <bondArray>         <bond atomRefs2="a1 a2" order="2" />         <bond atomRefs2="a1 a6" order="1" />         <bond atomRefs2="a2 a3" order="1" />         <bond atomRefs2="a3 a4" order="2" />         <bond atomRefs2="a4 a5" order="1" />         <bond atomRefs2="a5 a6" order="2" />         <bond atomRefs2="a1 a7" order="1" />         <bond atomRefs2="a7 a8" order="2" />         <bond atomRefs2="a6 a9" order="1" />         <bond atomRefs2="a8 a10" order="1" />         <bond atomRefs2="a10 a11" order="2" />         <bond atomRefs2="a10 a15" order="1" />         <bond atomRefs2="a11 a12" order="1" />         <bond atomRefs2="a12 a13" order="2" />         <bond atomRefs2="a13 a14" order="1" />         <bond atomRefs2="a14 a15" order="2" />       </bondArray>     </molecule>   </MChemicalStruct> </MDocument> </cml>

Antoher example would be the conjugated bond thing, polymer chemistry would be interested in (all-trans etc.):

Code:

<?xml version="1.0" ?> <cml> <MDocument>   <MChemicalStruct>     <molecule molID="m1">       <atomArray           atomID="a1 a2 a3 a4 a5 a6 a7"           elementType="C C C C C C C"           x2="-2.694999933242798 -1.3613208114147621 -0.02764168958672686 1.3060374322413084 1.3060374322413084 2.639716554069344 2.639716554069345"           y2="-9.239999771118164 -8.469999771118164 -9.239999771118164 -8.469999771118164 -6.9299997711181645 -6.159999771118165 -4.619999771118165"           />       <bondArray>         <bond atomRefs2="a1 a2" order="1" />         <bond atomRefs2="a2 a3" order="2" />         <bond atomRefs2="a3 a4" order="1" />         <bond atomRefs2="a4 a5" order="2" />         <bond atomRefs2="a5 a6" order="1" />         <bond atomRefs2="a6 a7" order="2" />       </bondArray>     </molecule>   </MChemicalStruct> </MDocument> </cml>

This molecule changes its stereochemistry obviously on applying the "Clean Up".














A shortcut to help peoply drawing nice molecules would be to implement a "grid"-like functions, where the molecules can snap in on rotation/translation etc..





But keep in mind, that the above examples are simple ones and usually one would be in a need for a "Beautify"-Function, when drawing big, complexes molecules (with stereoinformations included!).














btw: I like you CML-copy/paste-function very much!








:),





pheidrias

Hi,





Thank you for your comments about clean function.





As it turned out you need something else than we have it in clean 2D function.





The most important differences I see








- inhibit rotation around single bonds in conjugated system





- somehow take into account the actual coordinates and make the smallest change





As far as I know we don't have capacity to start this "beautify"-function.





What I can imagine to implement is to able to configure the drawing grid.





As you may have realized it is not possible to use any angle by simply adding new bond. 90 degree angle has 6 grid angle.


Would it be useful for you to able to configure the grid angle?


Such a way you would probably not need the "beautify"-function.





Andras

Hi,





regarding the alignment/rotation of groups, I think it's easy to do this without a grid, see the screenshots.





Regards, Hans-Ulrich

@volfi: This grid-functin would help a lot! Although I'm convinced, that many chemists would like to have this "beautify"-feature, as I described it. Honestly, this will be a crucial point in comparing your software with others...














@Hans-Ulrich (german?): Thanks at first for contributing (also with calculations - are you in theoretical chemistry business?)!





Please see the attached detail. At first I was only able to get it this way (an acceptable abberation), when rotating in 800%-magnification (in 100% I couldn't align it this near to vertical). But even here you can see on further magnification, that the vertical lines are not parallel to the screen - producing this staircase-effect.














Within this simple molecule it's of course not to much complicated to find a reasonable agreement between detail and time to put onto it, but please think on bigger molecules with many groups to be arranged!














Again: This "Beautify"-function would help you to get people attached to your software --> this 2D-drawing-program will be the attraction for people that lateron may want to buy other parts of software from your company!














Sincerely,





pheidrias

Hi Guntram,





I have a question about the "benzoate" in your first post of this topic, see attached screenshot.





Where is the negative charge ?








In this chemical formula I cannot recognize a chemical molecule or ion.





Using the tool "Markush Enumeration" shows 4 structures, but no benzoate.





What is the meaning of "to construct a ligand" ?





Regards, Hans-Ulrich

Hello Hans-Ulrich,





I didn't put the negative charge for easier building of a complex ligand ( http://www.chemaxon.com/forum/ftopic4527.html ).





This is btw. another good (bad) example for misleading 2D-structure-cleanup. When you clean up one of those molecules, the benoate function will point upward instead of down, as it has been drawn.





Regards,


pheidrias