ambit-users Mailing List for AMBIT:Chemical Structure DB&Web Service

Dear Yannick,

Concerning the stereo-chemistry:

Currently the code does not support the stereo-chemistry transformation 
(e.g. cis/trans double bonds, chirals atoms) due to the poor support of 
the stereo-chemistry in previous CDK versions (1.4.x) since we have just 
recently migrated to cdk 1.5.x. (this is ambit3)
So now the stereo-chemistry can be implemented fully on the base of new 
CDK 1.5.x.
It will take time to do it however. I expect in the following months to 
finish the stereo support.

With best regards
Nick




On 8/5/15 11:59 PM, Yannick .Djoumbou wrote:
> Hi all,
>
> I have an issue with the use of SMIRKS to transform molecules while 
> taking into account the stereochemistry. I have a smarts that depicts 
> a steroid analog carrying a 3-alpha hydroxyl group. I would like to 
> transform molecules matching this constraint, but I realized that 
> AMBIT is transforming any molecule that has the same skeleton depicted 
> in the reactant smarts independent of the stereochemistry (i.e. 
> 3-hydroxy, 3-alpha hydroxy, 3-beta hydroxy). Moreover, the 3D 
> arrangement of other chemical bonds is not conserved in the product.
>
> I use the SMIRKS:
> [#8;H1X2:1]-[#6;H1X4;@@:2]-1-[#6,#8,#7,#16:3]-,=[#6,#8,#7,#16:4]-,=[#6,#8,#7,#16:5]-,=2-,=[#6,#8,#7,#16;A;R2:6](-,=[#6,#8,#7,#16:7]-,=[#6,#8,#7,#16;A;R1:8]-,=[#6,#8,#7,#16;A;R2:9]-,=3-,=[#6,#8,#7,#16;A;R2:10]-,=4-,=[#6,#8,#7,#16;A;R1:11]-,=[#6,#8,#7,#16:12]-,=[#6,#8,#7,#16:13]-,=[#6,#8,#7,#16:14]-,=4-,=[#6,#8,#7,#16:15]-,=[#6,#8,#7,#16;A;R1:16]-,=[#6,#8,#7,#16:17]-,=2-,=3)-,=[#6,#8,#7,#16:18]-1>>[O;X1:1]=[#6;X3;@@:2]-1-[#6,#8,#7,#16:3]-,=[#6,#8,#7,#16:4]-,=[#6,#8,#7,#16:5]-,=2-,=[#6,#8,#7,#16;A;R2:6](-,=[#6,#8,#7,#16:7]-,=[#6,#8,#7,#16;A;R1:8]-,=[#6,#8,#7,#16;A;R2:9]-,=3-,=[#6,#8,#7,#16;A;R2:10]-,=4-,=[#6,#8,#7,#16;A;R1:11]-,=[#6,#8,#7,#16:12]-,=[#6,#8,#7,#16:13]-,=[#6,#8,#7,#16:14]-,=4-,=[#6,#8,#7,#16:15]-,=[#6,#8,#7,#16;A;R1:16]-,=[#6,#8,#7,#16:17]-,=2-,=3)-,=[#6,#8,#7,#16:18]-1
>
> And the function: applyTransformationWithSingleCopyForEachPos
>
> This is the smiles with the 3-alpha-hydroxyl group, the 3-beta-hydroxy 
> group, and a generic 3-hydroxyl group, respectively:
>
> A: CO[C@H]1[C@@H](O)C[C@H]2[C@@H]3CC=C4C[C@H](O)CC[C@]4(C)[C@H]3CC[C@]12C
>
> B: CO[C@H]1[C@@H](O)C[C@H]2[C@@H]3CC=C4C[C@@H](O)CC[C@]4(C)[C@H]3CC[C@]12C
>
> C: CO[C@H]1[C@@H](O)C[C@H]2[C@@H]3CC=C4CC(O)CC[C@]4(C)[C@H]3CC[C@]12C
>
> 1) Is it possible get AMBIT to transform only the 3-alpha-hydroxyl 
> bearing compounds based solely on the SMIRKS?  If yes, how?
>
> 2) I used a two step process where I first used CDK's SMARTSQueryTool 
> to do substructure matching and then transformed the molecules that 
> had a match.
>
> 3) One issue I had when using SMIRKS with or without prefiltering was 
> that the resulting structure did not contain any stereochemistry 
> information. The information stored in the input was lost throughout 
> the process.  How do I avoid this?
>
>
> Looking forward to hearing from you.
>
> Thank you for your consideration.
>
>
> Regards,
>
> Yannick

Hi Nina,

I have updated to the latest snapshot. I still have the same issue.
Hopefully, I will get some help soon.

Thank you.

Best,

Yannick.

On Wed, Aug 5, 2015 at 11:11 PM, Nina Jeliazkova <jel...@gm...>
wrote:

> Yannick,
>
> While waiting for Nick's answer (who's on vacation), please make sure you
> are using the latest 3.0.0-SNAPSHOT
>
>
> http://ambit.uni-plovdiv.bg:8083/nexus/#nexus-search;gav~ambit~ambit2-smarts~3.0.0-SNAPSHOT~~
>
> There were a number of SMIRKS-related updates during the last month, and
> judging from your other email you are using an older snapshot.
>
> Best regards,
> Nina
>
> On 5 August 2015 at 23:59, Yannick .Djoumbou <y.d...@gm...> wrote:
>
>> Hi all,
>>
>> I have an issue with the use of SMIRKS to transform molecules while
>> taking into account the stereochemistry. I have a smarts that depicts a
>> steroid analog carrying a 3-alpha hydroxyl group. I would like to transform
>> molecules matching this constraint, but I realized that AMBIT is
>> transforming any molecule that has the same skeleton depicted in the
>> reactant smarts independent of the stereochemistry (i.e. 3-hydroxy, 3-alpha
>> hydroxy, 3-beta hydroxy). Moreover, the 3D arrangement of other chemical
>> bonds is not conserved in the product.
>>
>> I use the SMIRKS:
>> [#8;H1X2:1]-[#6;H1X4;@
>> @:2]-1-[#6,#8,#7,#16:3]-,=[#6,#8,#7,#16:4]-,=[#6,#8,#7,#16:5]-,=2-,=[#6,#8,#7,#16;A;R2:6](-,=[#6,#8,#7,#16:7]-,=[#6,#8,#7,#16;A;R1:8]-,=[#6,#8,#7,#16;A;R2:9]-,=3-,=[#6,#8,#7,#16;A;R2:10]-,=4-,=[#6,#8,#7,#16;A;R1:11]-,=[#6,#8,#7,#16:12]-,=[#6,#8,#7,#16:13]-,=[#6,#8,#7,#16:14]-,=4-,=[#6,#8,#7,#16:15]-,=[#6,#8,#7,#16;A;R1:16]-,=[#6,#8,#7,#16:17]-,=2-,=3)-,=[#6,#8,#7,#16:18]-1>>[O;X1:1]=[#6;X3;@
>> @:2]-1-[#6,#8,#7,#16:3]-,=[#6,#8,#7,#16:4]-,=[#6,#8,#7,#16:5]-,=2-,=[#6,#8,#7,#16;A;R2:6](-,=[#6,#8,#7,#16:7]-,=[#6,#8,#7,#16;A;R1:8]-,=[#6,#8,#7,#16;A;R2:9]-,=3-,=[#6,#8,#7,#16;A;R2:10]-,=4-,=[#6,#8,#7,#16;A;R1:11]-,=[#6,#8,#7,#16:12]-,=[#6,#8,#7,#16:13]-,=[#6,#8,#7,#16:14]-,=4-,=[#6,#8,#7,#16:15]-,=[#6,#8,#7,#16;A;R1:16]-,=[#6,#8,#7,#16:17]-,=2-,=3)-,=[#6,#8,#7,#16:18]-1
>>
>> And the function: applyTransformationWithSingleCopyForEachPos
>>
>> This is the smiles with the 3-alpha-hydroxyl group, the 3-beta-hydroxy
>> group, and a generic 3-hydroxyl group, respectively:
>>
>> A:    CO[C@H]1[C@@H](O)C[C@H]2[C@@H]3CC=C4C[C@H](O)CC[C@]4(C)[C@H]3CC[C@
>> ]12C
>>
>> B:    CO[C@H]1[C@@H](O)C[C@H]2[C@@H]3CC=C4C[C@@H](O)CC[C@]4(C)[C@H]3CC[C@
>> ]12C
>>
>> C:    CO[C@H]1[C@@H](O)C[C@H]2[C@@H]3CC=C4CC(O)CC[C@]4(C)[C@H]3CC[C@]12C
>>
>> 1) Is it possible get AMBIT to transform only the 3-alpha-hydroxyl
>> bearing compounds based solely on the SMIRKS?  If yes, how?
>>
>> 2) I used a two step process where I first used CDK's SMARTSQueryTool to
>> do substructure matching and then transformed the molecules that had a
>> match.
>>
>> 3) One issue I had when using SMIRKS with or without prefiltering was
>> that the resulting structure did not contain any stereochemistry
>> information. The information stored in the input was lost throughout the
>> process.  How do I avoid this?
>>
>>
>> Looking forward to hearing from you.
>>
>> Thank you for your consideration.
>>
>>
>> Regards,
>>
>> Yannick
>>
>
>

Hi Nina,

Thanks for the reply. We were using the same version already. Well, I have
updated the snapshot version and had to change the code and use setter
functions instead. So, regarding this issue, it's all OK now.

Thank you.

Yannick

On Wed, Aug 5, 2015 at 10:26 PM, Nina Jeliazkova <jel...@gm...>
wrote:

> Yannick,
>
> Could you please ensure you are using the very same AMBIT snapshot .  My
> guess is the ones in MAC are using slightly outdated snapshot. Please note
> ambit.3.0.0-SNAPSHOT is a snapshot, not release, hence you should take care
> of working with the most recent one.
>
> I don't think this is OS issue (We are also using both MacOs and Windows).
>
> Regards,
> Nina
>
>
> On 6 August 2015 at 00:01, Yannick .Djoumbou <y.d...@gm...> wrote:
>
>>
>> ---------- Forwarded message ----------
>> From: Yannick .Djoumbou <y.d...@gm...>
>> Date: Fri, Jun 12, 2015 at 11:28 AM
>> Subject: SMIRKS MANAGER's Flags have to be set using different calls when
>> on Mac compared to Windows.
>> To: amb...@li...
>>
>>
>> Hi guys,
>>
>> I have been using the AMBIT library for a project. My colleague and I and
>> are developing a JAVA-based  application and are peer-programming on
>> Windows and MacOS X, respectively. We have not had any issue so far.
>> However, we realized that when trying to set the flags for the SMIRKS
>> Manager, we have to use a different from one platform to another.
>>
>> After creating a new SMIRKS manager:
>>    new SMIRKSManager(SilentChemObjectBuilder.getInstance());
>>
>> On MacOSX, we have to use:
>>      smrkMan.FlagFilterEquivalentMappings = true;
>>
>> On Windows, we have to use:
>>      smrkMan.setFlagProcessResultStructures(true);
>>
>> This is kind of troublesome since we are peer programming and have to
>> commit/push/pull the code that lies on the same repository.
>>
>> We are using the version 3.0.0 of AMBIT. How can we go around this?
>>
>> Thank you for your consideration.
>>
>> Regards,
>>
>> Yannick
>>
>>
>

Yannick,

While waiting for Nick's answer (who's on vacation), please make sure you
are using the latest 3.0.0-SNAPSHOT

http://ambit.uni-plovdiv.bg:8083/nexus/#nexus-search;gav~ambit~ambit2-smarts~3.0.0-SNAPSHOT~~

There were a number of SMIRKS-related updates during the last month, and
judging from your other email you are using an older snapshot.

Best regards,
Nina

On 5 August 2015 at 23:59, Yannick .Djoumbou <y.d...@gm...> wrote:

> Hi all,
>
> I have an issue with the use of SMIRKS to transform molecules while taking
> into account the stereochemistry. I have a smarts that depicts a steroid
> analog carrying a 3-alpha hydroxyl group. I would like to transform
> molecules matching this constraint, but I realized that AMBIT is
> transforming any molecule that has the same skeleton depicted in the
> reactant smarts independent of the stereochemistry (i.e. 3-hydroxy, 3-alpha
> hydroxy, 3-beta hydroxy). Moreover, the 3D arrangement of other chemical
> bonds is not conserved in the product.
>
> I use the SMIRKS:
> [#8;H1X2:1]-[#6;H1X4;@
> @:2]-1-[#6,#8,#7,#16:3]-,=[#6,#8,#7,#16:4]-,=[#6,#8,#7,#16:5]-,=2-,=[#6,#8,#7,#16;A;R2:6](-,=[#6,#8,#7,#16:7]-,=[#6,#8,#7,#16;A;R1:8]-,=[#6,#8,#7,#16;A;R2:9]-,=3-,=[#6,#8,#7,#16;A;R2:10]-,=4-,=[#6,#8,#7,#16;A;R1:11]-,=[#6,#8,#7,#16:12]-,=[#6,#8,#7,#16:13]-,=[#6,#8,#7,#16:14]-,=4-,=[#6,#8,#7,#16:15]-,=[#6,#8,#7,#16;A;R1:16]-,=[#6,#8,#7,#16:17]-,=2-,=3)-,=[#6,#8,#7,#16:18]-1>>[O;X1:1]=[#6;X3;@
> @:2]-1-[#6,#8,#7,#16:3]-,=[#6,#8,#7,#16:4]-,=[#6,#8,#7,#16:5]-,=2-,=[#6,#8,#7,#16;A;R2:6](-,=[#6,#8,#7,#16:7]-,=[#6,#8,#7,#16;A;R1:8]-,=[#6,#8,#7,#16;A;R2:9]-,=3-,=[#6,#8,#7,#16;A;R2:10]-,=4-,=[#6,#8,#7,#16;A;R1:11]-,=[#6,#8,#7,#16:12]-,=[#6,#8,#7,#16:13]-,=[#6,#8,#7,#16:14]-,=4-,=[#6,#8,#7,#16:15]-,=[#6,#8,#7,#16;A;R1:16]-,=[#6,#8,#7,#16:17]-,=2-,=3)-,=[#6,#8,#7,#16:18]-1
>
> And the function: applyTransformationWithSingleCopyForEachPos
>
> This is the smiles with the 3-alpha-hydroxyl group, the 3-beta-hydroxy
> group, and a generic 3-hydroxyl group, respectively:
>
> A:    CO[C@H]1[C@@H](O)C[C@H]2[C@@H]3CC=C4C[C@H](O)CC[C@]4(C)[C@H]3CC[C@
> ]12C
>
> B:    CO[C@H]1[C@@H](O)C[C@H]2[C@@H]3CC=C4C[C@@H](O)CC[C@]4(C)[C@H]3CC[C@
> ]12C
>
> C:    CO[C@H]1[C@@H](O)C[C@H]2[C@@H]3CC=C4CC(O)CC[C@]4(C)[C@H]3CC[C@]12C
>
> 1) Is it possible get AMBIT to transform only the 3-alpha-hydroxyl bearing
> compounds based solely on the SMIRKS?  If yes, how?
>
> 2) I used a two step process where I first used CDK's SMARTSQueryTool to
> do substructure matching and then transformed the molecules that had a
> match.
>
> 3) One issue I had when using SMIRKS with or without prefiltering was that
> the resulting structure did not contain any stereochemistry information.
> The information stored in the input was lost throughout the process.  How
> do I avoid this?
>
>
> Looking forward to hearing from you.
>
> Thank you for your consideration.
>
>
> Regards,
>
> Yannick
>

Yannick,

Could you please ensure you are using the very same AMBIT snapshot .  My
guess is the ones in MAC are using slightly outdated snapshot. Please note
ambit.3.0.0-SNAPSHOT is a snapshot, not release, hence you should take care
of working with the most recent one.

I don't think this is OS issue (We are also using both MacOs and Windows).

Regards,
Nina

On 6 August 2015 at 00:01, Yannick .Djoumbou <y.d...@gm...> wrote:

>
> ---------- Forwarded message ----------
> From: Yannick .Djoumbou <y.d...@gm...>
> Date: Fri, Jun 12, 2015 at 11:28 AM
> Subject: SMIRKS MANAGER's Flags have to be set using different calls when
> on Mac compared to Windows.
> To: amb...@li...
>
>
> Hi guys,
>
> I have been using the AMBIT library for a project. My colleague and I and
> are developing a JAVA-based  application and are peer-programming on
> Windows and MacOS X, respectively. We have not had any issue so far.
> However, we realized that when trying to set the flags for the SMIRKS
> Manager, we have to use a different from one platform to another.
>
> After creating a new SMIRKS manager:
>    new SMIRKSManager(SilentChemObjectBuilder.getInstance());
>
> On MacOSX, we have to use:
>      smrkMan.FlagFilterEquivalentMappings = true;
>
> On Windows, we have to use:
>      smrkMan.setFlagProcessResultStructures(true);
>
> This is kind of troublesome since we are peer programming and have to
> commit/push/pull the code that lies on the same repository.
>
> We are using the version 3.0.0 of AMBIT. How can we go around this?
>
> Thank you for your consideration.
>
> Regards,
>
> Yannick
>
>

---------- Forwarded message ----------
From: Yannick .Djoumbou <y.d...@gm...>
Date: Fri, Jun 12, 2015 at 11:28 AM
Subject: SMIRKS MANAGER's Flags have to be set using different calls when
on Mac compared to Windows.
To: amb...@li...


Hi guys,

I have been using the AMBIT library for a project. My colleague and I and
are developing a JAVA-based  application and are peer-programming on
Windows and MacOS X, respectively. We have not had any issue so far.
However, we realized that when trying to set the flags for the SMIRKS
Manager, we have to use a different from one platform to another.

After creating a new SMIRKS manager:
   new SMIRKSManager(SilentChemObjectBuilder.getInstance());

On MacOSX, we have to use:
     smrkMan.FlagFilterEquivalentMappings = true;

On Windows, we have to use:
     smrkMan.setFlagProcessResultStructures(true);

This is kind of troublesome since we are peer programming and have to
commit/push/pull the code that lies on the same repository.

We are using the version 3.0.0 of AMBIT. How can we go around this?

Thank you for your consideration.

Regards,

Yannick

Thank you Nina,

I still hope I will get an answer form Nick.

Regards,

Yannick

On Mon, Aug 3, 2015 at 7:20 AM, Nina Jeliazkova <jel...@gm...>
wrote:

> Hello Yannick,
>
> Nick might answer later this week, but generally, SMARTS are unable to
> express variable length chains, unless one do all the enumeration (which
> might be prohibitively large).
>
> Best regards,
> Nina
>
>
> On 30 July 2015 at 00:39, Yannick .Djoumbou <y.d...@gm...> wrote:
>
>> Hello,
>>
>> I would like to express the smarts for a class of compounds with an
>> aliphatic carbon chain of 13 to 21 carbon atoms, where  the longest chain
>> of carbon has between 13 and 21 carbons and is attached to a Nitrogen atom.
>> Examples include:
>>
>>     - CCCCCCCCCCCCCNH2 ; CCCCCC(C)C\C=C\C(C)CCCNH2 ;
>>  CCCCCCCCCCC\C=C\CCCCNH2
>>
>>  However, I do not want to include compounds containing a carbon chain of
>> more than 21 carbon atoms. Which means that the following compounds should
>> not be included:
>>
>>    -  CCCCCCCCCCCCCCCC\C=C\CCCCNH2; CC/C=C\CCCCCCCCCCCC\C=C\CCCCNH2
>>
>>
>> I first generated the patterns for each of the two types:
>>
>> 13 carbons attached to N: (A)  =>
>>  [#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-[#7]([H])[H]
>>
>> 22 carbons attached to N: (B)  =>
>>  [#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-[#7]([H])[H]
>>
>> I then tried to combined those into  1 smarts [$(A)!$(B)] to mean that
>> the compound of interest should match A but not match B. This does not
>> work!
>>
>> What would be the best way of describing this in one SMARTS string?
>>
>> Thank you
>>
>> Regards,
>>
>> MrYan
>>
>
>

Hi all,

I have an issue with the use of SMIRKS to transform molecules while taking
into account the stereochemistry. I have a smarts that depicts a steroid
analog carrying a 3-alpha hydroxyl group. I would like to transform
molecules matching this constraint, but I realized that AMBIT is
transforming any molecule that has the same skeleton depicted in the
reactant smarts independent of the stereochemistry (i.e. 3-hydroxy, 3-alpha
hydroxy, 3-beta hydroxy). Moreover, the 3D arrangement of other chemical
bonds is not conserved in the product.

I use the SMIRKS:
[#8;H1X2:1]-[#6;H1X4;@
@:2]-1-[#6,#8,#7,#16:3]-,=[#6,#8,#7,#16:4]-,=[#6,#8,#7,#16:5]-,=2-,=[#6,#8,#7,#16;A;R2:6](-,=[#6,#8,#7,#16:7]-,=[#6,#8,#7,#16;A;R1:8]-,=[#6,#8,#7,#16;A;R2:9]-,=3-,=[#6,#8,#7,#16;A;R2:10]-,=4-,=[#6,#8,#7,#16;A;R1:11]-,=[#6,#8,#7,#16:12]-,=[#6,#8,#7,#16:13]-,=[#6,#8,#7,#16:14]-,=4-,=[#6,#8,#7,#16:15]-,=[#6,#8,#7,#16;A;R1:16]-,=[#6,#8,#7,#16:17]-,=2-,=3)-,=[#6,#8,#7,#16:18]-1>>[O;X1:1]=[#6;X3;@
@:2]-1-[#6,#8,#7,#16:3]-,=[#6,#8,#7,#16:4]-,=[#6,#8,#7,#16:5]-,=2-,=[#6,#8,#7,#16;A;R2:6](-,=[#6,#8,#7,#16:7]-,=[#6,#8,#7,#16;A;R1:8]-,=[#6,#8,#7,#16;A;R2:9]-,=3-,=[#6,#8,#7,#16;A;R2:10]-,=4-,=[#6,#8,#7,#16;A;R1:11]-,=[#6,#8,#7,#16:12]-,=[#6,#8,#7,#16:13]-,=[#6,#8,#7,#16:14]-,=4-,=[#6,#8,#7,#16:15]-,=[#6,#8,#7,#16;A;R1:16]-,=[#6,#8,#7,#16:17]-,=2-,=3)-,=[#6,#8,#7,#16:18]-1

And the function: applyTransformationWithSingleCopyForEachPos

This is the smiles with the 3-alpha-hydroxyl group, the 3-beta-hydroxy
group, and a generic 3-hydroxyl group, respectively:

A:    CO[C@H]1[C@@H](O)C[C@H]2[C@@H]3CC=C4C[C@H](O)CC[C@]4(C)[C@H]3CC[C@]12C

B:    CO[C@H]1[C@@H](O)C[C@H]2[C@@H]3CC=C4C[C@@H](O)CC[C@]4(C)[C@H]3CC[C@
]12C

C:    CO[C@H]1[C@@H](O)C[C@H]2[C@@H]3CC=C4CC(O)CC[C@]4(C)[C@H]3CC[C@]12C

1) Is it possible get AMBIT to transform only the 3-alpha-hydroxyl bearing
compounds based solely on the SMIRKS?  If yes, how?

2) I used a two step process where I first used CDK's SMARTSQueryTool to do
substructure matching and then transformed the molecules that had a match.

3) One issue I had when using SMIRKS with or without prefiltering was that
the resulting structure did not contain any stereochemistry information.
The information stored in the input was lost throughout the process.  How
do I avoid this?


Looking forward to hearing from you.

Thank you for your consideration.


Regards,

Yannick

Hello Yannick,

Nick might answer later this week, but generally, SMARTS are unable to
express variable length chains, unless one do all the enumeration (which
might be prohibitively large).

Best regards,
Nina

On 30 July 2015 at 00:39, Yannick .Djoumbou <y.d...@gm...> wrote:

> Hello,
>
> I would like to express the smarts for a class of compounds with an
> aliphatic carbon chain of 13 to 21 carbon atoms, where  the longest chain
> of carbon has between 13 and 21 carbons and is attached to a Nitrogen atom.
> Examples include:
>
>     - CCCCCCCCCCCCCNH2 ; CCCCCC(C)C\C=C\C(C)CCCNH2 ;
>  CCCCCCCCCCC\C=C\CCCCNH2
>
>  However, I do not want to include compounds containing a carbon chain of
> more than 21 carbon atoms. Which means that the following compounds should
> not be included:
>
>    -  CCCCCCCCCCCCCCCC\C=C\CCCCNH2; CC/C=C\CCCCCCCCCCCC\C=C\CCCCNH2
>
>
> I first generated the patterns for each of the two types:
>
> 13 carbons attached to N: (A)  =>
>  [#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-[#7]([H])[H]
>
> 22 carbons attached to N: (B)  =>
>  [#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-,=[#6]-[#7]([H])[H]
>
> I then tried to combined those into  1 smarts [$(A)!$(B)] to mean that the
> compound of interest should match A but not match B. This does not work!
>
> What would be the best way of describing this in one SMARTS string?
>
> Thank you
>
> Regards,
>
> MrYan
>