{
  "id": 94991,
  "title": "Structure Coordinates and Orientation",
  "url": "/competitions/champs-scalar-coupling/discussion/94991",
  "author_name": "",
  "post_date": "2019-06-09T02:29:17.441722100Z",
  "votes": 1,
  "comment_count": 1,
  "views": 0,
  "content": "<p>I did attempt to find this information in the competition overview, data, and discussions. Maybe a domain expert can sweep in here with a short summary or some easily digested links.</p>\n\n<p>On what basis are the molecules positioned and oriented? It's not obvious and seems arbitrary. If I kept digging I might eventually discover this through EDA, but why suffer?</p>\n\n<p>Irrelevant, newbie follow-up question: What are the units? They are presumably to the same scale every time?</p>",
  "messages": [
    {
      "id": "548233",
      "postDate": "06/09/2019 02:29:17",
      "content": "<p>I did attempt to find this information in the competition overview, data, and discussions. Maybe a domain expert can sweep in here with a short summary or some easily digested links.</p>\n\n<p>On what basis are the molecules positioned and oriented? It's not obvious and seems arbitrary. If I kept digging I might eventually discover this through EDA, but why suffer?</p>\n\n<p>Irrelevant, newbie follow-up question: What are the units? They are presumably to the same scale every time?</p>",
      "rawMarkdown": "I did attempt to find this information in the competition overview, data, and discussions. Maybe a domain expert can sweep in here with a short summary or some easily digested links.\n\nOn what basis are the molecules positioned and oriented? It's not obvious and seems arbitrary. If I kept digging I might eventually discover this through EDA, but why suffer?\n\nIrrelevant, newbie follow-up question: What are the units? They are presumably to the same scale every time?",
      "votes": null
    },
    {
      "id": "548237",
      "postDate": "06/09/2019 02:37:19",
      "content": "<p>Not sure about the orientation. I feel like creating a rotation-invariant model is ideal for this reason. You could also try to align each molecule along it's principal components and place the centroid on the origin to get a standardized orientation.</p>\n\n<p>And yes, they are to scale and the units is Angstroms (0.1 nm).</p>",
      "rawMarkdown": "Not sure about the orientation. I feel like creating a rotation-invariant model is ideal for this reason. You could also try to align each molecule along it's principal components and place the centroid on the origin to get a standardized orientation.\n\nAnd yes, they are to scale and the units is Angstroms (0.1 nm).",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 548237,
      "author_name": "matthewmasters",
      "author_url": "",
      "post_date": "06/09/2019 02:37:19",
      "content": "<p>Not sure about the orientation. I feel like creating a rotation-invariant model is ideal for this reason. You could also try to align each molecule along it's principal components and place the centroid on the origin to get a standardized orientation.</p>\n\n<p>And yes, they are to scale and the units is Angstroms (0.1 nm).</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "548233": "I did attempt to find this information in the competition overview, data, and discussions. Maybe a domain expert can sweep in here with a short summary or some easily digested links.\n\nOn what basis are the molecules positioned and oriented? It's not obvious and seems arbitrary. If I kept digging I might eventually discover this through EDA, but why suffer?\n\nIrrelevant, newbie follow-up question: What are the units? They are presumably to the same scale every time?",
    "548237": "Not sure about the orientation. I feel like creating a rotation-invariant model is ideal for this reason. You could also try to align each molecule along it's principal components and place the centroid on the origin to get a standardized orientation.\n\nAnd yes, they are to scale and the units is Angstroms (0.1 nm)."
  },
  "source": "meta"
}