For those of you without the domain expertise, it helps to understand the difference between \"direct\" and \"indirect\" couplings between atomic nuclear dipole moments of a molecule.
\nThe \"direct\" coupling is a through-space dipole-dipole interaction, while the \"indirect\" or scalar coupling is an electron-mediated through-bond interaction.\nTypically the \"direct\" dipole coupling of molecules in solution (in an NMR spectrometer) is averaged out due to rapid thermal molecular rotation in comparison to the timescale of the NMR experiment and only the scalar coupling remains.\nI should know, I got my PH.D in solid-state NMR studies of indirect couplings.
For those of you without the domain expertise, it helps to understand the difference between \"direct\" and \"indirect\" couplings between atomic nuclear dipole moments of a molecule.
\nThe \"direct\" coupling is a through-space dipole-dipole interaction, while the \"indirect\" or scalar coupling is an electron-mediated through-bond interaction.\nTypically the \"direct\" dipole coupling of molecules in solution (in an NMR spectrometer) is averaged out due to rapid thermal molecular rotation in comparison to the timescale of the NMR experiment and only the scalar coupling remains.\nI should know, I got my PH.D in solid-state NMR studies of indirect couplings.
Thanks for the heads up. Now I just need to Google what that means.
", "rawMarkdown": "Thanks for the heads up. Now I just need to Google what that means.", "votes": null }, { "id": "539892", "postDate": "05/30/2019 16:58:56", "content": "> JaredSavage wrote\n> \n> > Thanks for the heads up. Now I just need to Google what that means.\n.
\n\nExactly, if possible, please provide a more intuitive answer, or point out the literature required.
", "rawMarkdown": "> **JaredSavage wrote**\n> \n> > Thanks for the heads up. Now I just need to Google what that means.\n.\n\n\n\n\nExactly, if possible, please provide a more intuitive answer, or point out the literature required.", "votes": null }, { "id": "539912", "postDate": "05/30/2019 17:36:12", "content": "It means that models that use simple geometry to calculate couplings (how far each carbon is from each hydrogen atom, or what is the angle between 3 atoms) is unlikely to capture the underlying physical/chemical process. Instead an additional layer of knowledge may be helpful. For example, if a carbon and hydrogen are 1.06 Å, then it is likely a certain class of chemical bond which might have a characteristic coupling constant. Or, as another example, if two atoms are a spatially close together, but they are not connected by < 4 chemical bonds, they may not have any significant coupling. Does that help give an idea of the flavor of what the OP was saying?
", "rawMarkdown": "It means that models that use simple geometry to calculate couplings (how far each carbon is from each hydrogen atom, or what is the angle between 3 atoms) is unlikely to capture the underlying physical/chemical process. Instead an additional layer of knowledge may be helpful. For example, if a carbon and hydrogen are 1.06 Å, then it is likely a certain class of chemical bond which might have a characteristic coupling constant. Or, as another example, if two atoms are a spatially close together, but they are not connected by < 4 chemical bonds, they may not have any significant coupling. Does that help give an idea of the flavor of what the OP was saying?", "votes": null }, { "id": "540001", "postDate": "05/30/2019 20:56:01", "content": "Yes, that is quite enlightening. Thank you very much.
", "rawMarkdown": "Yes, that is quite enlightening. Thank you very much.", "votes": null }, { "id": "540454", "postDate": "05/31/2019 14:42:16", "content": "It is important to notice that even though molecular structure optimization methods result in indicating specific potential energy minimum, in many cases molecular structures are not rigid. Very often they undergo rotations around chemical bonds resulting in attaining conformations different than the one indicated as a global minimum. In such cases measured molecular properties, such as NMR chemical shifts and coupling constants will be an average. Of course, the effect of conformational averaging wil be an issue only if transition barriers are low given the temperature at which the spectra were measured, and also if global minimum is not separated from other local minima by too large gibbs energy differences.
\n\nHowever, values provided in the competition seem to be results of theoretical calculations anyway, hence, they were obtained from \"frozen\" structures, and the above is not an issue.
", "rawMarkdown": "It is important to notice that even though molecular structure optimization methods result in indicating specific potential energy minimum, in many cases molecular structures are not rigid. Very often they undergo rotations around chemical bonds resulting in attaining conformations different than the one indicated as a global minimum. In such cases measured molecular properties, such as NMR chemical shifts and coupling constants will be an average. Of course, the effect of conformational averaging wil be an issue only if transition barriers are low given the temperature at which the spectra were measured, and also if global minimum is not separated from other local minima by too large gibbs energy differences.\n\nHowever, values provided in the competition seem to be results of theoretical calculations anyway, hence, they were obtained from \"frozen\" structures, and the above is not an issue.", "votes": null }, { "id": "540469", "postDate": "05/31/2019 15:01:20", "content": "I've noticed that... the structures seem to be computer-generated at 0 K. So at the end of the competition, CHAMPS will have a very accurate model to predict J-couplings for theoretical molecules, but it won't be much practical use when you measure average J-couplings at room temperature.
", "rawMarkdown": "I've noticed that... the structures seem to be computer-generated at 0 K. So at the end of the competition, CHAMPS will have a very accurate model to predict J-couplings for theoretical molecules, but it won't be much practical use when you measure average J-couplings at room temperature.", "votes": null }, { "id": "540547", "postDate": "05/31/2019 16:09:22", "content": "\n", "rawMarkdown": "> if a carbon and hydrogen are 1.06 Å\nWould it mean C is sp3 hybridized? Aren't bond lengths for sp3 and sp2 too close together to be confused in strained molecules? Wouldn't the number of bonds or angles be a better predictor? I am a total noob, operating on my high-school knowledge here, so please bear with me :)", "votes": null }, { "id": "540594", "postDate": "05/31/2019 17:45:38", "content": "if a carbon and hydrogen are 1.06 Å\n Would it mean C is sp3 hybridized? Aren't bond lengths for sp3 and sp2 too close together to be confused in strained molecules? Wouldn't the number of bonds or angles be a better predictor? I am a total noob, operating on my high-school knowledge here, so please bear with me :)
\n
Yeah, I don't think we can tell the hybridisation from the C-H bond lengths which are 1.09, 1.08 and 1.07 for sp3, sp2 and sp.
", "rawMarkdown": "Yeah, I don't think we can tell the hybridisation from the C-H bond lengths which are 1.09, 1.08 and 1.07 for sp3, sp2 and sp.", "votes": null }, { "id": "541746", "postDate": "06/03/2019 01:33:03", "content": "", "rawMarkdown": "", "votes": null }, { "id": "541748", "postDate": "06/03/2019 01:33:35", "content": "Are you saying the coupling constants in the dataset are not empirical coupling constants measured by NMR, but the results of some theoretical calculation with multiple unrealistic (zero Kelvin, no solvent, DFT, etc. etc.) assumptions?
", "rawMarkdown": "Are you saying the coupling constants in the dataset are not empirical coupling constants measured by NMR, but the results of some theoretical calculation with multiple unrealistic (zero Kelvin, no solvent, DFT, etc. etc.) assumptions?", "votes": null }, { "id": "541866", "postDate": "06/03/2019 06:11:26", "content": "Exactly
", "rawMarkdown": "Exactly", "votes": null } ], "comments": [ { "id": 539710, "author_name": "jaredsavage", "author_url": "", "post_date": "05/30/2019 12:30:58", "content": "Thanks for the heads up. Now I just need to Google what that means.
", "votes": null, "replies": [ { "id": 539892, "author_name": "chazarnik", "author_url": "", "post_date": "05/30/2019 16:58:56", "content": "> JaredSavage wrote\n> \n> > Thanks for the heads up. Now I just need to Google what that means.\n.
\n\nExactly, if possible, please provide a more intuitive answer, or point out the literature required.
", "votes": null, "replies": [] }, { "id": 539912, "author_name": "jjohnsumn", "author_url": "", "post_date": "05/30/2019 17:36:12", "content": "It means that models that use simple geometry to calculate couplings (how far each carbon is from each hydrogen atom, or what is the angle between 3 atoms) is unlikely to capture the underlying physical/chemical process. Instead an additional layer of knowledge may be helpful. For example, if a carbon and hydrogen are 1.06 Å, then it is likely a certain class of chemical bond which might have a characteristic coupling constant. Or, as another example, if two atoms are a spatially close together, but they are not connected by < 4 chemical bonds, they may not have any significant coupling. Does that help give an idea of the flavor of what the OP was saying?
", "votes": null, "replies": [] }, { "id": 540001, "author_name": "chazarnik", "author_url": "", "post_date": "05/30/2019 20:56:01", "content": "Yes, that is quite enlightening. Thank you very much.
", "votes": null, "replies": [] }, { "id": 540547, "author_name": "andris", "author_url": "", "post_date": "05/31/2019 16:09:22", "content": "\n", "votes": null, "replies": [] }, { "id": 540594, "author_name": "phucco", "author_url": "", "post_date": "05/31/2019 17:45:38", "content": "if a carbon and hydrogen are 1.06 Å\n Would it mean C is sp3 hybridized? Aren't bond lengths for sp3 and sp2 too close together to be confused in strained molecules? Wouldn't the number of bonds or angles be a better predictor? I am a total noob, operating on my high-school knowledge here, so please bear with me :)
\n
Yeah, I don't think we can tell the hybridisation from the C-H bond lengths which are 1.09, 1.08 and 1.07 for sp3, sp2 and sp.
", "votes": null, "replies": [] } ] }, { "id": 540454, "author_name": "wojciechmigda", "author_url": "", "post_date": "05/31/2019 14:42:16", "content": "It is important to notice that even though molecular structure optimization methods result in indicating specific potential energy minimum, in many cases molecular structures are not rigid. Very often they undergo rotations around chemical bonds resulting in attaining conformations different than the one indicated as a global minimum. In such cases measured molecular properties, such as NMR chemical shifts and coupling constants will be an average. Of course, the effect of conformational averaging wil be an issue only if transition barriers are low given the temperature at which the spectra were measured, and also if global minimum is not separated from other local minima by too large gibbs energy differences.
\n\nHowever, values provided in the competition seem to be results of theoretical calculations anyway, hence, they were obtained from \"frozen\" structures, and the above is not an issue.
", "votes": null, "replies": [ { "id": 540469, "author_name": "phucco", "author_url": "", "post_date": "05/31/2019 15:01:20", "content": "I've noticed that... the structures seem to be computer-generated at 0 K. So at the end of the competition, CHAMPS will have a very accurate model to predict J-couplings for theoretical molecules, but it won't be much practical use when you measure average J-couplings at room temperature.
", "votes": null, "replies": [] }, { "id": 541748, "author_name": "degraw", "author_url": "", "post_date": "06/03/2019 01:33:35", "content": "Are you saying the coupling constants in the dataset are not empirical coupling constants measured by NMR, but the results of some theoretical calculation with multiple unrealistic (zero Kelvin, no solvent, DFT, etc. etc.) assumptions?
", "votes": null, "replies": [] }, { "id": 541866, "author_name": "phucco", "author_url": "", "post_date": "06/03/2019 06:11:26", "content": "Exactly
", "votes": null, "replies": [] } ] }, { "id": 541746, "author_name": "degraw", "author_url": "", "post_date": "06/03/2019 01:33:03", "content": "", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "539581": "For those of you without the domain expertise, it helps to understand the difference between \"direct\" and \"indirect\" couplings between atomic nuclear dipole moments of a molecule. \nThe \"direct\" coupling is a through-space dipole-dipole interaction, while the \"indirect\" or scalar coupling is an electron-mediated through-bond interaction.\nTypically the \"direct\" dipole coupling of molecules in solution (in an NMR spectrometer) is averaged out due to rapid thermal molecular rotation in comparison to the timescale of the NMR experiment and only the scalar coupling remains.\nI should know, I got my PH.D in solid-state NMR studies of indirect couplings.", "539710": "Thanks for the heads up. Now I just need to Google what that means.", "539892": "> **JaredSavage wrote**\n> \n> > Thanks for the heads up. Now I just need to Google what that means.\n.\n\n\n\n\nExactly, if possible, please provide a more intuitive answer, or point out the literature required.", "539912": "It means that models that use simple geometry to calculate couplings (how far each carbon is from each hydrogen atom, or what is the angle between 3 atoms) is unlikely to capture the underlying physical/chemical process. Instead an additional layer of knowledge may be helpful. For example, if a carbon and hydrogen are 1.06 Å, then it is likely a certain class of chemical bond which might have a characteristic coupling constant. Or, as another example, if two atoms are a spatially close together, but they are not connected by < 4 chemical bonds, they may not have any significant coupling. Does that help give an idea of the flavor of what the OP was saying?", "540001": "Yes, that is quite enlightening. Thank you very much.", "540454": "It is important to notice that even though molecular structure optimization methods result in indicating specific potential energy minimum, in many cases molecular structures are not rigid. Very often they undergo rotations around chemical bonds resulting in attaining conformations different than the one indicated as a global minimum. In such cases measured molecular properties, such as NMR chemical shifts and coupling constants will be an average. Of course, the effect of conformational averaging wil be an issue only if transition barriers are low given the temperature at which the spectra were measured, and also if global minimum is not separated from other local minima by too large gibbs energy differences.\n\nHowever, values provided in the competition seem to be results of theoretical calculations anyway, hence, they were obtained from \"frozen\" structures, and the above is not an issue.", "540469": "I've noticed that... the structures seem to be computer-generated at 0 K. So at the end of the competition, CHAMPS will have a very accurate model to predict J-couplings for theoretical molecules, but it won't be much practical use when you measure average J-couplings at room temperature.", "540547": "> if a carbon and hydrogen are 1.06 Å\nWould it mean C is sp3 hybridized? Aren't bond lengths for sp3 and sp2 too close together to be confused in strained molecules? Wouldn't the number of bonds or angles be a better predictor? I am a total noob, operating on my high-school knowledge here, so please bear with me :)", "540594": "Yeah, I don't think we can tell the hybridisation from the C-H bond lengths which are 1.09, 1.08 and 1.07 for sp3, sp2 and sp.", "541746": "", "541748": "Are you saying the coupling constants in the dataset are not empirical coupling constants measured by NMR, but the results of some theoretical calculation with multiple unrealistic (zero Kelvin, no solvent, DFT, etc. etc.) assumptions?", "541866": "Exactly" }, "source": "meta" }