search

TinkerTools / tinker

Tinker: Software Tools for Molecular Design
https://dasher.wustl.edu/tinker/
Other
130 stars 61 forks source link

Bad interpolation in molecule with triple bond linear angles with minimize program #91

Closed misterbrandonwalker closed 3 years ago

misterbrandonwalker commented 3 years ago

Hey Dr. Ponder,

I have a molecule with a triple bond, kinda like a stick off the ring. It seems even without any torsion restraints the minimizer has hard time finding "complete convergence". Seems to get stuck in knowing right direction (maybe due to undefined dihedral plane?). Just wondering if this is like something that is "fixable" internally. I know for QM like Gaussian opt, they have trouble with this in modified redundant coordinates but sometimes can fix with cartesian coordinates instead. Hoping for any suggestions! 3561.key_4.txt 3561.xyz_2.txt

misterbrandonwalker commented 3 years ago

Okay Dr. Ren suggest zero out torsion on colinear and that works

jayponder commented 3 years ago

Hi Brandon,

Yes, exactly. I just did the same thing and the minimization works fine. You must use "zero" torsion parameters on any torsion where either three or all four of the atoms involved are colinear. Those torsion angle values are numerically ill-defined as the value of the angle changes wildly with tiny changes in atomic coordinates.

By the way, if you have Tinker questions issues like this, or the one yesterday, I might prefer that you just email me directly instead of using Github.

Best, Jay

misterbrandonwalker commented 3 years ago

Sounds good thanks!

On Thu, Jul 1, 2021 at 11:36 AM Jay Ponder @.***> wrote:

Hi Brandon,

Yes, exactly. I just did the same thing and the minimization works fine. You must use "zero" torsion parameters on any torsion where either three or all four of the atoms involved are colinear. Those torsion angle values are numerically ill-defined as the value of the angle changes wildly with tiny changes in atomic coordinates.

By the way, if you have Tinker questions issues like this, or the one yesterday, I might prefer that you just email me directly instead of using Github.

Best, Jay

— You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub https://github.com/TinkerTools/tinker/issues/91#issuecomment-872392414, or unsubscribe https://github.com/notifications/unsubscribe-auth/AKNB26NF42DNZTVWBICDYK3TVSKSFANCNFSM47VBKZEQ .

zjing7 commented 3 years ago

Hi Jay,

I think it would be a good idea to have a user forum for questions like this, so that next time people have similar questions, they can search for answers, or someone may answer the question for us.

Best, Francis

On Thu, Jul 1, 2021 at 11:36 AM Jay Ponder @.***> wrote:

Hi Brandon,

Yes, exactly. I just did the same thing and the minimization works fine. You must use "zero" torsion parameters on any torsion where either three or all four of the atoms involved are colinear. Those torsion angle values are numerically ill-defined as the value of the angle changes wildly with tiny changes in atomic coordinates.

By the way, if you have Tinker questions issues like this, or the one yesterday, I might prefer that you just email me directly instead of using Github.

Best, Jay

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/TinkerTools/tinker/issues/91#issuecomment-872392414, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABW6OE42W3K2RI4OSYPERXTTVSKSHANCNFSM47VBKZEQ .

-- Zhifeng (Francis) Jing

jayponder commented 3 years ago

Hi Francis,

Yes, that's the intent of the TinkerTools.org website (https://www.TinkerTools.org), which has a Forum built into the site. Some time ago I got the basic site put in place, but I've never filled it out to make it really useful. At some point, we will need to "advertise" this site and its Forum, but I think we are not quite ready yet. The Forum requires people to register on the site (when I first set it up, without registration, lots of people "spammed" the Forum...).

Best, Jay

zjing7 commented 3 years ago

I didn't know about this. Thanks!

On Thu, Jul 1, 2021 at 12:21 PM Jay Ponder @.***> wrote:

Hi Francis,

Yes, that's the intent of the TinkerTools.org website ( https://www.TinkerTools.org), which has a Forum built into the site. Some time ago I got the basic site put in place, but I've never filled it out to make it really useful. At some point, we will need to "advertise" this site and its Forum, but I think we are not quite ready yet. The Forum requires people to register on the site (when I first set it up, without registration, lots of people "spammed" the Forum...).

Best, Jay

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/TinkerTools/tinker/issues/91#issuecomment-872421153, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABW6OE6U25C3OJYDIB7VJIDTVSP2LANCNFSM47VBKZEQ .

-- Zhifeng (Francis) Jing