{
  "id": 680551,
  "title": "Chain ordering",
  "url": "/competitions/stanford-rna-3d-folding-2/discussion/680551",
  "author_name": "Gabor Balazs",
  "post_date": "2026-03-09T08:47:58.288000",
  "votes": 3,
  "comment_count": 2,
  "views": 0,
  "content": "<p>In a multi-chain RNA sequence, how are the chains ordered? Do they follow some kind of common consensus, a biological constraint maybe, or is it random (i.e., chosen by authors based on what they think important)?</p>",
  "messages": [
    {
      "id": 3419686,
      "postDate": "2026-03-11T11:54:50.847Z",
      "content": "<p>In datasets, chain ordering is essentially author-defined as there's no enforced standard.\nLoose conventions exist (largest chain first, RNA before protein, functional order), but two depositions of the same complex can differ. For ML pipelines this matters because your model shouldn't assume chain order carries consistent meaning across samples.</p>",
      "rawMarkdown": "In datasets, chain ordering is essentially author-defined as there's no enforced standard.\nLoose conventions exist (largest chain first, RNA before protein, functional order), but two depositions of the same complex can differ. For ML pipelines this matters because your model shouldn't assume chain order carries consistent meaning across samples.",
      "votes": 3
    },
    {
      "id": 3418788,
      "postDate": "2026-03-09T08:47:58.287Z",
      "content": "<p>In a multi-chain RNA sequence, how are the chains ordered? Do they follow some kind of common consensus, a biological constraint maybe, or is it random (i.e., chosen by authors based on what they think important)?</p>",
      "rawMarkdown": "In a multi-chain RNA sequence, how are the chains ordered? Do they follow some kind of common consensus, a biological constraint maybe, or is it random (i.e., chosen by authors based on what they think important)?",
      "votes": 3
    },
    {
      "id": 3419809,
      "postDate": "2026-03-11T18:30:56.220Z",
      "content": "<p>In the <code>*_sequences.csv</code> files the order of the chains in <code>stoichiometry</code> is arbitrary. The <code>sequence</code> follows the order in the <code>stoichiometry</code>. For example, if there are three chains, two with a sequence  <code>A=GCGC</code> and one with a sequence <code>B=AUAU</code>, then the <code>stoichiometry</code> can be either <code>A:2;B:1</code> or <code>B:1;A:2</code>. That will correspond to the concatenated sequence <code>GCGCGCGCAUAU</code> or <code>AUAUGCGCGCGC</code> respectively. </p>",
      "rawMarkdown": "In the `*_sequences.csv` files the order of the chains in `stoichiometry` is arbitrary. The `sequence` follows the order in the `stoichiometry`. For example, if there are three chains, two with a sequence  `A=GCGC` and one with a sequence `B=AUAU`, then the `stoichiometry` can be either `A:2;B:1` or `B:1;A:2`. That will correspond to the concatenated sequence `GCGCGCGCAUAU` or `AUAUGCGCGCGC` respectively. ",
      "votes": 2
    }
  ],
  "comments": [
    {
      "id": 3419686,
      "author_name": "Manpreet Kaur",
      "author_url": "",
      "post_date": "2026-03-11T11:54:50.847000",
      "content": "<p>In datasets, chain ordering is essentially author-defined as there's no enforced standard.\nLoose conventions exist (largest chain first, RNA before protein, functional order), but two depositions of the same complex can differ. For ML pipelines this matters because your model shouldn't assume chain order carries consistent meaning across samples.</p>",
      "votes": 3,
      "replies": []
    },
    {
      "id": 3419809,
      "author_name": "Przemek Porebski",
      "author_url": "",
      "post_date": "2026-03-11T18:30:56.220000",
      "content": "<p>In the <code>*_sequences.csv</code> files the order of the chains in <code>stoichiometry</code> is arbitrary. The <code>sequence</code> follows the order in the <code>stoichiometry</code>. For example, if there are three chains, two with a sequence  <code>A=GCGC</code> and one with a sequence <code>B=AUAU</code>, then the <code>stoichiometry</code> can be either <code>A:2;B:1</code> or <code>B:1;A:2</code>. That will correspond to the concatenated sequence <code>GCGCGCGCAUAU</code> or <code>AUAUGCGCGCGC</code> respectively. </p>",
      "votes": 2,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "3419686": "In datasets, chain ordering is essentially author-defined as there's no enforced standard.\nLoose conventions exist (largest chain first, RNA before protein, functional order), but two depositions of the same complex can differ. For ML pipelines this matters because your model shouldn't assume chain order carries consistent meaning across samples.",
    "3418788": "In a multi-chain RNA sequence, how are the chains ordered? Do they follow some kind of common consensus, a biological constraint maybe, or is it random (i.e., chosen by authors based on what they think important)?",
    "3419809": "In the `*_sequences.csv` files the order of the chains in `stoichiometry` is arbitrary. The `sequence` follows the order in the `stoichiometry`. For example, if there are three chains, two with a sequence  `A=GCGC` and one with a sequence `B=AUAU`, then the `stoichiometry` can be either `A:2;B:1` or `B:1;A:2`. That will correspond to the concatenated sequence `GCGCGCGCAUAU` or `AUAUGCGCGCGC` respectively. "
  }
}