{
  "id": 546417,
  "title": "Can we assume fully assambled ribosomes?",
  "url": "/competitions/czii-cryo-et-object-identification/discussion/546417",
  "author_name": "",
  "post_date": "2024-11-15T16:14:39.619711800Z",
  "votes": 2,
  "comment_count": 4,
  "views": 0,
  "content": "<p>The assembly of ribosomes requires factors (proteins) that are present at lower concentrations in this \"phantom\" dataset compared to cells.</p>\n<p>Yet, it seems to be assumed that they are mostly assembled.</p>\n<p>Is the fraction of partly assembled ribosomes negligible?<br>\nIf not, were the partly assembled ribosomes annotated as ribosomes or not?</p>",
  "messages": [
    {
      "id": "3046601",
      "postDate": "11/15/2024 16:14:39",
      "content": "<p>The assembly of ribosomes requires factors (proteins) that are present at lower concentrations in this \"phantom\" dataset compared to cells.</p>\n<p>Yet, it seems to be assumed that they are mostly assembled.</p>\n<p>Is the fraction of partly assembled ribosomes negligible?<br>\nIf not, were the partly assembled ribosomes annotated as ribosomes or not?</p>",
      "rawMarkdown": "The assembly of ribosomes requires factors (proteins) that are present at lower concentrations in this \"phantom\" dataset compared to cells.\n\nYet, it seems to be assumed that they are mostly assembled.\n\nIs the fraction of partly assembled ribosomes negligible?\nIf not, were the partly assembled ribosomes annotated as ribosomes or not?",
      "votes": null
    },
    {
      "id": "3046869",
      "postDate": "11/16/2024 01:47:27",
      "content": "<p>Hi Gyula, great question. The annotated ribosomes are fully assembled 80s ribosomes. The dataset also has large subunit and small subunit ribosomes but those are not part of the challenge.</p>",
      "rawMarkdown": "Hi Gyula, great question. The annotated ribosomes are fully assembled 80s ribosomes. The dataset also has large subunit and small subunit ribosomes but those are not part of the challenge.",
      "votes": null
    },
    {
      "id": "3047341",
      "postDate": "11/16/2024 15:19:46",
      "content": "<p>Hi,<br>\nThank you for the answer!</p>\n<p>Can you give a rough estimate of what percentage is assambled?</p>\n<p>How confident are you that you did not annote any unessambled subunits as ribosomes and did not miss any of the assambled ribosomes thinking it was just a large subunit?</p>\n<p>I would assume it is not that easy to differntiate between fully assambled ribosomes and the large subunit.</p>",
      "rawMarkdown": "Hi,\nThank you for the answer!\n\nCan you give a rough estimate of what percentage is assambled?\n\nHow confident are you that you did not annote any unessambled subunits as ribosomes and did not miss any of the assambled ribosomes thinking it was just a large subunit?\n\nI would assume it is not that easy to differntiate between fully assambled ribosomes and the large subunit.",
      "votes": null
    },
    {
      "id": "3048304",
      "postDate": "11/17/2024 18:24:14",
      "content": "<p>We did 2D and 3D averages of the annotated ribosomes. We chose the subset that reconstructed to a full ribosome. The current ground truth has some false negatives, in other words, we missed some of the full ribosomes. I have an idea of what the full/subunit ratio might be, but since I'm not certain I rather not mislead you. It always comes down to the problem that it's incredibly difficult to confidently annotate all the ground truth in the data. So what I can say is that we are confident about the annotated full ribosomes that are part of the scoring, and we acknowledge that we missed some of the full ribosomes. This is part of the reason the beta value for scoring was set to 4, so that we don't severely punish true positives that are deemed false positive with respect to our ground truth. So 16 false positives are punished as much as 1 false negative.</p>",
      "rawMarkdown": "We did 2D and 3D averages of the annotated ribosomes. We chose the subset that reconstructed to a full ribosome. The current ground truth has some false negatives, in other words, we missed some of the full ribosomes. I have an idea of what the full/subunit ratio might be, but since I'm not certain I rather not mislead you. It always comes down to the problem that it's incredibly difficult to confidently annotate all the ground truth in the data. So what I can say is that we are confident about the annotated full ribosomes that are part of the scoring, and we acknowledge that we missed some of the full ribosomes. This is part of the reason the beta value for scoring was set to 4, so that we don't severely punish true positives that are deemed false positive with respect to our ground truth. So 16 false positives are punished as much as 1 false negative.",
      "votes": null
    },
    {
      "id": "3048324",
      "postDate": "11/17/2024 18:47:29",
      "content": "<p>Thank you very much for the detailed answer!</p>",
      "rawMarkdown": "Thank you very much for the detailed answer!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3046869,
      "author_name": "rezaparaan",
      "author_url": "",
      "post_date": "11/16/2024 01:47:27",
      "content": "<p>Hi Gyula, great question. The annotated ribosomes are fully assembled 80s ribosomes. The dataset also has large subunit and small subunit ribosomes but those are not part of the challenge.</p>",
      "votes": null,
      "replies": [
        {
          "id": 3047341,
          "author_name": "gyulamaloveczky4",
          "author_url": "",
          "post_date": "11/16/2024 15:19:46",
          "content": "<p>Hi,<br>\nThank you for the answer!</p>\n<p>Can you give a rough estimate of what percentage is assambled?</p>\n<p>How confident are you that you did not annote any unessambled subunits as ribosomes and did not miss any of the assambled ribosomes thinking it was just a large subunit?</p>\n<p>I would assume it is not that easy to differntiate between fully assambled ribosomes and the large subunit.</p>",
          "votes": null,
          "replies": [
            {
              "id": 3048304,
              "author_name": "rezaparaan",
              "author_url": "",
              "post_date": "11/17/2024 18:24:14",
              "content": "<p>We did 2D and 3D averages of the annotated ribosomes. We chose the subset that reconstructed to a full ribosome. The current ground truth has some false negatives, in other words, we missed some of the full ribosomes. I have an idea of what the full/subunit ratio might be, but since I'm not certain I rather not mislead you. It always comes down to the problem that it's incredibly difficult to confidently annotate all the ground truth in the data. So what I can say is that we are confident about the annotated full ribosomes that are part of the scoring, and we acknowledge that we missed some of the full ribosomes. This is part of the reason the beta value for scoring was set to 4, so that we don't severely punish true positives that are deemed false positive with respect to our ground truth. So 16 false positives are punished as much as 1 false negative.</p>",
              "votes": null,
              "replies": [
                {
                  "id": 3048324,
                  "author_name": "gyulamaloveczky4",
                  "author_url": "",
                  "post_date": "11/17/2024 18:47:29",
                  "content": "<p>Thank you very much for the detailed answer!</p>",
                  "votes": null,
                  "replies": []
                }
              ]
            }
          ]
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "3046601": "The assembly of ribosomes requires factors (proteins) that are present at lower concentrations in this \"phantom\" dataset compared to cells.\n\nYet, it seems to be assumed that they are mostly assembled.\n\nIs the fraction of partly assembled ribosomes negligible?\nIf not, were the partly assembled ribosomes annotated as ribosomes or not?",
    "3046869": "Hi Gyula, great question. The annotated ribosomes are fully assembled 80s ribosomes. The dataset also has large subunit and small subunit ribosomes but those are not part of the challenge.",
    "3047341": "Hi,\nThank you for the answer!\n\nCan you give a rough estimate of what percentage is assambled?\n\nHow confident are you that you did not annote any unessambled subunits as ribosomes and did not miss any of the assambled ribosomes thinking it was just a large subunit?\n\nI would assume it is not that easy to differntiate between fully assambled ribosomes and the large subunit.",
    "3048304": "We did 2D and 3D averages of the annotated ribosomes. We chose the subset that reconstructed to a full ribosome. The current ground truth has some false negatives, in other words, we missed some of the full ribosomes. I have an idea of what the full/subunit ratio might be, but since I'm not certain I rather not mislead you. It always comes down to the problem that it's incredibly difficult to confidently annotate all the ground truth in the data. So what I can say is that we are confident about the annotated full ribosomes that are part of the scoring, and we acknowledge that we missed some of the full ribosomes. This is part of the reason the beta value for scoring was set to 4, so that we don't severely punish true positives that are deemed false positive with respect to our ground truth. So 16 false positives are punished as much as 1 false negative.",
    "3048324": "Thank you very much for the detailed answer!"
  },
  "source": "meta"
}