Torsions of small non-aromatic rings and quarternary carbon atoms in rings

[D16ERG]

Dear all,

During the parameterization of delta 24 Ergosterol, I noticed that this molecule suffers from missing torsions. For this reason I experimented with smaller, comparable systems. Through the SMARTS keys, I noticed that quarternary carbons atoms in (fused) rings results in missing torsions. While bare cycloalkanes display a somewhat similar issue; I don't believe this was the cause of these missing torsions. IIRC the latter issue was caused by an inactive fragmenter (<25 atoms).

On another note, I would like to ask what skills it takes to properly grasp Poltype. Currently I am finishing my BSc. in general chemistry. Herein I did follow the quantum and physical chemistry courses. As a reference I currently own "The Theory of Intermolecular Forces" by Anthony Stone. In addition to this I do realize Poltype heavily relies on Python. Despite this; Poltype seems to be particularly challenging to grasp, let alone develop. (this latter comment is not meant in any bad way)

Kind regards,

Didier

[misterbrandonwalker]

Hello,

Please attach a link of all input and output files and I can take a look. For now can just read original poltype publications and the online documentation and let us know if anything is confusing.

• Brandon

[D16ERG]

Dear Brandon,

All input and output of the latter molecule have been compressed into the following zip-file: https://wetransfer.com/downloads/6742f6cf8be571e278e3b59a21ba13cd20210729174117/f1f8d7

Kind regards,

Didier

[misterbrandonwalker]

When I click on the link it in the website location it just says " http://url/" and says this site cannot be reached.

On Thu, Jul 29, 2021 at 12:44 PM D16ERG @.***> wrote:

Dear Brandon,

All input and output of the latter molecule have been compressed into the following zip-file: https://we.tl/t-7K3pdgzT5a http://url

Kind regards,

Didier

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/pren/poltype/issues/95#issuecomment-889337825, or unsubscribe https://github.com/notifications/unsubscribe-auth/AKNB26OGM3EJLE7J7QSWP5DT2GHOXANCNFSM5BG5FFEA .

[D16ERG]

I just updated the link. Regarding what I don't understand is the following example:

Transferring from ['Alkane -H2C-', 'Alkane -CH2-', 'Alkane -CH2-', 'Alkane -H2C-']

Fitted from Fragment SMARTS #6(-#6(-[H])-[H])(-[H])-[H] torsion atom indexes = 7,2,0,4 with smarts torsion indices 12,6,1,14 from fragment 13_1_Index_0.mol

torsion % #6(-#6(-[H])-[H])(-[H])-[H] % 12,6,1,14 % 0,0,0.299

I do grasp the SMART format of describing molecules. What I did not understand were the torsion atom indices and the smarts atom indices. Especially the "0" in this excerpt "torsion atom indexes = 7,2,0,4 " is confusing.

[leucinw]

I believe you should get rid of https://

On Thu, Jul 29, 2021 at 12:48 PM Brandon Walker @.***> wrote:

When I click on the link it in the website location it just says " http://url/" and says this site cannot be reached.

On Thu, Jul 29, 2021 at 12:44 PM D16ERG @.***> wrote:

Dear Brandon,

All input and output of the latter molecule have been compressed into the following zip-file: https://we.tl/t-7K3pdgzT5a http://url

Kind regards,

Didier

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/pren/poltype/issues/95#issuecomment-889337825, or unsubscribe < https://github.com/notifications/unsubscribe-auth/AKNB26OGM3EJLE7J7QSWP5DT2GHOXANCNFSM5BG5FFEA

.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/pren/poltype/issues/95#issuecomment-889340551, or unsubscribe https://github.com/notifications/unsubscribe-auth/AHI63NLPJ7JZWO4MTYEHNSDT2GH7HANCNFSM5BG5FFEA .

[pren]

https://we.tl/t-7K3pdgzT5a this worked for me

From: Brandon Walker @.> Sent: Thursday, July 29, 2021 12:49 PM To: pren/poltype @.> Cc: Subscribed @.***> Subject: Re: [pren/poltype] Torsions of small non-aromatic rings and quarternary carbon atoms in rings (#95)

When I click on the link it in the website location it just says " http://url/" and says this site cannot be reached.

On Thu, Jul 29, 2021 at 12:44 PM D16ERG @.<mailto:@.>> wrote:

Dear Brandon,

All input and output of the latter molecule have been compressed into the following zip-file: https://we.tl/t-7K3pdgzT5a http://url

Kind regards,

Didier

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/pren/poltype/issues/95#issuecomment-889337825, or unsubscribe https://github.com/notifications/unsubscribe-auth/AKNB26OGM3EJLE7J7QSWP5DT2GHOXANCNFSM5BG5FFEA .

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHubhttps://github.com/pren/poltype/issues/95#issuecomment-889340551, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ABNC6XUJ464A522EBPYABM3T2GH7HANCNFSM5BG5FFEA.

[misterbrandonwalker]

Yes thank you the first time did not work but after updating the link works fine.

Fitted from Fragment SMARTS

#6(-#6(-[H])-[H])(-[H])-[H] torsion atom indexes = 7,2,0,4 with smarts torsion indices 12,6,1,14 from fragment 13_1_Index_0.mol

torsion %

#6(-#6(-[H])-[H])(-[H])-[H] % 12,6,1,14 % 0,0,0.299 torsion 415 401 405 413 0 0.0 1 0 180.0 2 0.299 0.0 3

Torsion atom indexes are like if you open molecule in pymol and start atoms with 0 based index, which atoms are in torsion. Smarts torsion indices are like which atoms in smart string counting left to right starting from 1.

On Thu, Jul 29, 2021 at 1:05 PM Pengyu Ren @.***> wrote:

https://we.tl/t-7K3pdgzT5a this worked for me

From: Brandon Walker @.> Sent: Thursday, July 29, 2021 12:49 PM To: pren/poltype @.> Cc: Subscribed @.***> Subject: Re: [pren/poltype] Torsions of small non-aromatic rings and quarternary carbon atoms in rings (#95)

When I click on the link it in the website location it just says " http://url/" and says this site cannot be reached.

On Thu, Jul 29, 2021 at 12:44 PM D16ERG @.<mailto:@.>> wrote:

Dear Brandon,

All input and output of the latter molecule have been compressed into the following zip-file: https://we.tl/t-7K3pdgzT5a http://url

Kind regards,

Didier

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/pren/poltype/issues/95#issuecomment-889337825, or unsubscribe < https://github.com/notifications/unsubscribe-auth/AKNB26OGM3EJLE7J7QSWP5DT2GHOXANCNFSM5BG5FFEA>

.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub< https://github.com/pren/poltype/issues/95#issuecomment-889340551>, or unsubscribe< https://github.com/notifications/unsubscribe-auth/ABNC6XUJ464A522EBPYABM3T2GH7HANCNFSM5BG5FFEA>.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/pren/poltype/issues/95#issuecomment-889351285, or unsubscribe https://github.com/notifications/unsubscribe-auth/AKNB26JJJEOGCOIWBO6XEJ3T2GJ65ANCNFSM5BG5FFEA .

[misterbrandonwalker]

So I know it will say in poltype.log parameters are missing, then will fit or assign default values but your final key file is not "missing" any parameters. Can check this via analyze tinker program with xyz and key file.

On Thu, Jul 29, 2021 at 1:07 PM Brandon Walker @.***> wrote:

Yes thank you the first time did not work but after updating the link works fine.

Fitted from Fragment SMARTS

#6(-#6(-[H])-[H])(-[H])-[H] torsion atom indexes = 7,2,0,4 with smarts torsion indices 12,6,1,14 from fragment 13_1_Index_0.mol

torsion %

#6(-#6(-[H])-[H])(-[H])-[H] % 12,6,1,14 % 0,0,0.299 torsion 415 401 405 413 0 0.0 1 0 180.0 2 0.299 0.0 3

Torsion atom indexes are like if you open molecule in pymol and start atoms with 0 based index, which atoms are in torsion. Smarts torsion indices are like which atoms in smart string counting left to right starting from 1.

On Thu, Jul 29, 2021 at 1:05 PM Pengyu Ren @.***> wrote:

https://we.tl/t-7K3pdgzT5a this worked for me

From: Brandon Walker @.> Sent: Thursday, July 29, 2021 12:49 PM To: pren/poltype @.> Cc: Subscribed @.***> Subject: Re: [pren/poltype] Torsions of small non-aromatic rings and quarternary carbon atoms in rings (#95)

When I click on the link it in the website location it just says " http://url/" and says this site cannot be reached.

On Thu, Jul 29, 2021 at 12:44 PM D16ERG @.<mailto:@.>> wrote:

Dear Brandon,

All input and output of the latter molecule have been compressed into the following zip-file: https://we.tl/t-7K3pdgzT5a http://url

Kind regards,

Didier

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/pren/poltype/issues/95#issuecomment-889337825, or unsubscribe <

https://github.com/notifications/unsubscribe-auth/AKNB26OGM3EJLE7J7QSWP5DT2GHOXANCNFSM5BG5FFEA

.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub< https://github.com/pren/poltype/issues/95#issuecomment-889340551>, or unsubscribe<

https://github.com/notifications/unsubscribe-auth/ABNC6XUJ464A522EBPYABM3T2GH7HANCNFSM5BG5FFEA .

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/pren/poltype/issues/95#issuecomment-889351285, or unsubscribe < https://github.com/notifications/unsubscribe-auth/AKNB26JJJEOGCOIWBO6XEJ3T2GJ65ANCNFSM5BG5FFEA

.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/pren/poltype/issues/95#issuecomment-889352689, or unsubscribe https://github.com/notifications/unsubscribe-auth/AKNB26PHZJZDRVBF7V5A35LT2GKG7ANCNFSM5BG5FFEA .

[D16ERG]

Could I have a faulty installation? When I restart poltype from the directory I just sent; I only receive a ValueError:

_Poltype -- Polarizable atom typer of small molecules for the AMOEBA polarizable force field

Please cite: Wu, J.C.; Chattree, G.; Ren, P.Y.; Automation of AMOEBA polarizable force field parameterization for small molecules. Theor Chem Acc.

Version 2.3.1 May 2021

Copyright (c) Johnny Wu, Gaurav Chattree, Brandon Walker, Matthew Harger and Pengyu Ren 2019

All Rights Reserved

ValueError: torsion parameters are all zero torsion 401 407 402 409 0 0.0 1 0 180.0 2 0 0.0 3 path =/home/didier/Desktop/ABring_

[misterbrandonwalker]

Thank you, the error you see gets piped to standard output and not into the poltype.log so I would not have seen it without running myself. The issue is rather complicated to explain, there is a section in the docs (databaseparser.py) regarding treatment of torsion on a ring and what is assigned via an array of hybridizations (SP3/SP2 etc) for each atom of the torsion, there have been some "special" cases I have run across involving torsion between rings that complicates things. Our rules are supposed to decide if the ring is "non-aromatic enough" for its own fragment or if needs to be transferred torsion from alkane/benzene. Easy case is all 4 atoms on 1 ring, another case is like in adenine-diphosphate, where lets say 3 atoms are on sugar ring (SP3), but last atom is SP2 on adenine, then if you just use your array of hybridizations following the rules in the docs you would not be able to make its own fragment for torsion parameters, so what I have done previously is essentially only include the atoms on that first ring where the rotatable bond is. In your case, is also a special case where there is non-aromatic ring attached to mostly aromatic. Essentially I believe the program just saw the first two on non-aromatic ring and then thinks its able to make its own fragment, however this fragment was not made ( i have another part of program that checks for which torsions are viable for fragments), then since the parameters were set to 0 in first key file (set to 0 in order to fit), but not fitting occured then there is an error at the end...

Anyway, for now I have temporarily changed it to not remove any atoms in the hybdrization array and now alkane parameters will be transferred since the middle bond has one SP3 and one SP2. Just update program and delete key_4 and .key_5 and restart.

1. If b aromatic and c not aromatic or (b non-aromatic and c aromatic).
   1. If H-X-X-X transfer from H-C-C-C
   2. If H-X-X-H transfer from H-C-C-H
   3. If X-X-X-X transfer from C-C-C-C

On Thu, Jul 29, 2021 at 1:15 PM D16ERG @.***> wrote:

Could I have a faulty installation? When I restart poltype from the directory I just sent I only receive a ValueError:

Poltype -- Polarizable atom typer of small molecules for the AMOEBA polarizable force field

Please cite: Wu, J.C.; Chattree, G.; Ren, P.Y.; Automation of AMOEBA polarizable force field parameterization for small molecules. Theor Chem Acc.

Version 2.3.1 May 2021

Copyright (c) Johnny Wu, Gaurav Chattree, Brandon Walker, Matthew Harger and Pengyu Ren 2019

All Rights Reserved

ValueError: torsion parameters are all zero torsion 401 407 402 409 0 0.0 1 0 180.0 2 0 0.0 3 path =/home/didier/Desktop/ABring

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/pren/poltype/issues/95#issuecomment-889357474, or unsubscribe https://github.com/notifications/unsubscribe-auth/AKNB26LLKWZPUZC4VYUOIN3T2GLEJANCNFSM5BG5FFEA .

[D16ERG]

Just read your comment. The updated code is running now.

[D16ERG]

Dear Brandon and Pengyu,

Thank for your time and efforts! The parameterization of the above fragment ran just fine. I tried piecing together a solution with the current extensive documentation. Unfortunately this is well above my pay grade.

For those who are interested: the parameterizations will be incorporated into a QM/MM setup. This setup features the progams Autodock, Tinker, Poltype2, BAGEL and Newton-X to simulate the photochemistry of a host-guest complex which features a vitamin D. Since the molecules involved are moderately big; not every atom is involved in the actual photochemistry of the excited states. Hence the use of QM with BAGEL and MM with Tinker.

This setup was inspired by the following paper: Catalyzed Thermal Isomerization between Previtamin D3 and Vitamin D3 via β-Cyclodextrin Complexation DOI:https://doi.org/10.1074/jbc.270.15.8706

Kind regards,

Didier