{
  "id": 94340,
  "title": "what is the preferred library basis to used for spin coupling calculation?",
  "url": "/competitions/champs-scalar-coupling/discussion/94340",
  "author_name": "",
  "post_date": "2019-06-04T02:30:26.369434200Z",
  "votes": 2,
  "comment_count": 4,
  "views": 0,
  "content": "<p>i can find a list of basis library here:</p>\n\n<p><a href=\"https://svn.pnl.gov/svn/nwchem/trunk/src/basis/libraries/\">https://svn.pnl.gov/svn/nwchem/trunk/src/basis/libraries/</a></p>\n\n<p>i read that  cc-pvtz is the preferred one, is it true?</p>",
  "messages": [
    {
      "id": "542634",
      "postDate": "06/04/2019 02:30:26",
      "content": "<p>i can find a list of basis library here:</p>\n\n<p><a href=\"https://svn.pnl.gov/svn/nwchem/trunk/src/basis/libraries/\">https://svn.pnl.gov/svn/nwchem/trunk/src/basis/libraries/</a></p>\n\n<p>i read that  cc-pvtz is the preferred one, is it true?</p>",
      "rawMarkdown": "i can find a list of basis library here:\n\nhttps://svn.pnl.gov/svn/nwchem/trunk/src/basis/libraries/\n\ni read that  cc-pvtz is the preferred one, is it true?",
      "votes": null
    },
    {
      "id": "542652",
      "postDate": "06/04/2019 02:46:14",
      "content": "<p>also, i find that 1JHC of methane is about 121 , e.g. </p>\n\n<p><a href=\"https://core.ac.uk/download/pdf/52478121.pdf\">https://core.ac.uk/download/pdf/52478121.pdf</a>. (Computational NMR coupling constants: Shifting and scaling factors for evaluating 1JCH)\n<img src=\"https://storage.googleapis.com/kaggle-forum-message-attachments/542652/13380/ch4.png\" alt=\"\">\nBut the value given by kaggle data  (dsgdb9nsd_000001 ) is only 84.8?\n<code>\n  id,molecule_name,atom_index_0,atom_index_1,type,scalar_coupling_constant\n  0,dsgdb9nsd_000001,1,0,1JHC,84.8076\n</code>\nWhy is this so?</p>",
      "rawMarkdown": "also, i find that 1JHC of methane is about 121 , e.g. \n\nhttps://core.ac.uk/download/pdf/52478121.pdf. (Computational NMR coupling constants: Shifting and scaling factors for evaluating 1JCH)\n![](https://storage.googleapis.com/kaggle-forum-message-attachments/542652/13380/ch4.png)\nBut the value given by kaggle data  (dsgdb9nsd_000001 ) is only 84.8?\n```\n  id,molecule_name,atom_index_0,atom_index_1,type,scalar_coupling_constant\n  0,dsgdb9nsd_000001,1,0,1JHC,84.8076\n```\nWhy is this so?",
      "votes": null
    },
    {
      "id": "543634",
      "postDate": "06/04/2019 16:43:21",
      "content": "<p>I think it's because these structures are the result of a QM simulation at 0 K, so it doesn't reflect reality at higher temperatures, or include dynamics.</p>\n\n<p>Experimentally the 1JHC coupling is 125 Hz (see abstract of <a href=\"https://www.sciencedirect.com/science/article/pii/0022236491900755\">https://www.sciencedirect.com/science/article/pii/0022236491900755</a>), not 85 Hz like we're given.</p>\n\n<p>The bond lengths in the structure file are 1.092 A, whereas they should be over 1.1 A, once you \"relax\" the structure (I'm using CrystalMaker) or look at other calculations (<a href=\"http://www.colby.edu/chemistry/webmo/CH4.html\">http://www.colby.edu/chemistry/webmo/CH4.html</a>)</p>",
      "rawMarkdown": "I think it's because these structures are the result of a QM simulation at 0 K, so it doesn't reflect reality at higher temperatures, or include dynamics.\n\nExperimentally the 1JHC coupling is 125 Hz (see abstract of https://www.sciencedirect.com/science/article/pii/0022236491900755), not 85 Hz like we're given.\n\nThe bond lengths in the structure file are 1.092 A, whereas they should be over 1.1 A, once you \"relax\" the structure (I'm using CrystalMaker) or look at other calculations (http://www.colby.edu/chemistry/webmo/CH4.html)",
      "votes": null
    },
    {
      "id": "555463",
      "postDate": "06/18/2019 23:09:29",
      "content": "<p>Actually I've just checked some textbook and found that the C-H bond in methane is about 1.09 A, so perhaps the structures are optimized. It's still odd why the coupling constant is inaccurate. I hope the host can clarify this.</p>",
      "rawMarkdown": "Actually I've just checked some textbook and found that the C-H bond in methane is about 1.09 A, so perhaps the structures are optimized. It's still odd why the coupling constant is inaccurate. I hope the host can clarify this.",
      "votes": null
    },
    {
      "id": "555626",
      "postDate": "06/19/2019 07:03:20",
      "content": "<p>Good spot. I'm not sure why Crystalmaker relaxes to 1.1 and that Colby website also reports 1.1.\nThe structures aren't relaxed though, you can see that from the hydrogen positions in some of the bigger/strained molecules.\nMy guess is that the hosts will take the best ML models and then train them on either experimental or more accurately computed J-couplings.</p>",
      "rawMarkdown": "Good spot. I'm not sure why Crystalmaker relaxes to 1.1 and that Colby website also reports 1.1.\nThe structures aren't relaxed though, you can see that from the hydrogen positions in some of the bigger/strained molecules.\nMy guess is that the hosts will take the best ML models and then train them on either experimental or more accurately computed J-couplings.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 542652,
      "author_name": "hengck23",
      "author_url": "",
      "post_date": "06/04/2019 02:46:14",
      "content": "<p>also, i find that 1JHC of methane is about 121 , e.g. </p>\n\n<p><a href=\"https://core.ac.uk/download/pdf/52478121.pdf\">https://core.ac.uk/download/pdf/52478121.pdf</a>. (Computational NMR coupling constants: Shifting and scaling factors for evaluating 1JCH)\n<img src=\"https://storage.googleapis.com/kaggle-forum-message-attachments/542652/13380/ch4.png\" alt=\"\">\nBut the value given by kaggle data  (dsgdb9nsd_000001 ) is only 84.8?\n<code>\n  id,molecule_name,atom_index_0,atom_index_1,type,scalar_coupling_constant\n  0,dsgdb9nsd_000001,1,0,1JHC,84.8076\n</code>\nWhy is this so?</p>",
      "votes": null,
      "replies": [
        {
          "id": 543634,
          "author_name": "phucco",
          "author_url": "",
          "post_date": "06/04/2019 16:43:21",
          "content": "<p>I think it's because these structures are the result of a QM simulation at 0 K, so it doesn't reflect reality at higher temperatures, or include dynamics.</p>\n\n<p>Experimentally the 1JHC coupling is 125 Hz (see abstract of <a href=\"https://www.sciencedirect.com/science/article/pii/0022236491900755\">https://www.sciencedirect.com/science/article/pii/0022236491900755</a>), not 85 Hz like we're given.</p>\n\n<p>The bond lengths in the structure file are 1.092 A, whereas they should be over 1.1 A, once you \"relax\" the structure (I'm using CrystalMaker) or look at other calculations (<a href=\"http://www.colby.edu/chemistry/webmo/CH4.html\">http://www.colby.edu/chemistry/webmo/CH4.html</a>)</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 555463,
          "author_name": "mdlhvn",
          "author_url": "",
          "post_date": "06/18/2019 23:09:29",
          "content": "<p>Actually I've just checked some textbook and found that the C-H bond in methane is about 1.09 A, so perhaps the structures are optimized. It's still odd why the coupling constant is inaccurate. I hope the host can clarify this.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 555626,
          "author_name": "phucco",
          "author_url": "",
          "post_date": "06/19/2019 07:03:20",
          "content": "<p>Good spot. I'm not sure why Crystalmaker relaxes to 1.1 and that Colby website also reports 1.1.\nThe structures aren't relaxed though, you can see that from the hydrogen positions in some of the bigger/strained molecules.\nMy guess is that the hosts will take the best ML models and then train them on either experimental or more accurately computed J-couplings.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "542634": "i can find a list of basis library here:\n\nhttps://svn.pnl.gov/svn/nwchem/trunk/src/basis/libraries/\n\ni read that  cc-pvtz is the preferred one, is it true?",
    "542652": "also, i find that 1JHC of methane is about 121 , e.g. \n\nhttps://core.ac.uk/download/pdf/52478121.pdf. (Computational NMR coupling constants: Shifting and scaling factors for evaluating 1JCH)\n![](https://storage.googleapis.com/kaggle-forum-message-attachments/542652/13380/ch4.png)\nBut the value given by kaggle data  (dsgdb9nsd_000001 ) is only 84.8?\n```\n  id,molecule_name,atom_index_0,atom_index_1,type,scalar_coupling_constant\n  0,dsgdb9nsd_000001,1,0,1JHC,84.8076\n```\nWhy is this so?",
    "543634": "I think it's because these structures are the result of a QM simulation at 0 K, so it doesn't reflect reality at higher temperatures, or include dynamics.\n\nExperimentally the 1JHC coupling is 125 Hz (see abstract of https://www.sciencedirect.com/science/article/pii/0022236491900755), not 85 Hz like we're given.\n\nThe bond lengths in the structure file are 1.092 A, whereas they should be over 1.1 A, once you \"relax\" the structure (I'm using CrystalMaker) or look at other calculations (http://www.colby.edu/chemistry/webmo/CH4.html)",
    "555463": "Actually I've just checked some textbook and found that the C-H bond in methane is about 1.09 A, so perhaps the structures are optimized. It's still odd why the coupling constant is inaccurate. I hope the host can clarify this.",
    "555626": "Good spot. I'm not sure why Crystalmaker relaxes to 1.1 and that Colby website also reports 1.1.\nThe structures aren't relaxed though, you can see that from the hydrogen positions in some of the bigger/strained molecules.\nMy guess is that the hosts will take the best ML models and then train them on either experimental or more accurately computed J-couplings."
  },
  "source": "meta"
}