{
  "id": 347086,
  "title": "Cite-seq test data cell type label released by mistake?",
  "url": "/competitions/open-problems-multimodal/discussion/347086",
  "author_name": "",
  "post_date": "2022-08-22T19:03:21.900020700Z",
  "votes": 6,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Hello there,</p>\n<p>I was doing some EDA today with the metadata.csv.</p>\n<p>While trying to summarize the cell type information, I found that the Cite-seq test data label is available while 10XMultiome ones were hidden. Is it supposed to be so or by mistake?</p>\n<p>Here is the plot:<br>\n<img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4455340%2F99638aac0d2205b27ee5e41648053430%2FPicture1.png?generation=1661194992446561&amp;alt=media\" alt=\"\"></p>",
  "messages": [
    {
      "id": "1909624",
      "postDate": "08/22/2022 19:03:21",
      "content": "<p>Hello there,</p>\n<p>I was doing some EDA today with the metadata.csv.</p>\n<p>While trying to summarize the cell type information, I found that the Cite-seq test data label is available while 10XMultiome ones were hidden. Is it supposed to be so or by mistake?</p>\n<p>Here is the plot:<br>\n<img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4455340%2F99638aac0d2205b27ee5e41648053430%2FPicture1.png?generation=1661194992446561&amp;alt=media\" alt=\"\"></p>",
      "rawMarkdown": "Hello there,\n\nI was doing some EDA today with the metadata.csv.\n\nWhile trying to summarize the cell type information, I found that the Cite-seq test data label is available while 10XMultiome ones were hidden. Is it supposed to be so or by mistake?\n\nHere is the plot:\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4455340%2F99638aac0d2205b27ee5e41648053430%2FPicture1.png?generation=1661194992446561&alt=media)",
      "votes": null
    },
    {
      "id": "1909697",
      "postDate": "08/22/2022 20:55:31",
      "content": "<p>Hi <a href=\"https://www.kaggle.com/comp540cl123\" target=\"_blank\">@comp540cl123</a>! Good catch! This is intended: we derive cell type from RNA data. Therefore, to avoid any leak of information, we do not release RNA for the data set where RNA is the target for the prediction and therefore hidden  (as the case for the Multiome). For the CITEseq data, the RNA for the test is released - as it serves as the input for the predictions for surface protein expressions. Therefore, we released it there. Does that answer your question?</p>",
      "rawMarkdown": "Hi @comp540cl123! Good catch! This is intended: we derive cell type from RNA data. Therefore, to avoid any leak of information, we do not release RNA for the data set where RNA is the target for the prediction and therefore hidden  (as the case for the Multiome). For the CITEseq data, the RNA for the test is released - as it serves as the input for the predictions for surface protein expressions. Therefore, we released it there. Does that answer your question?",
      "votes": null
    },
    {
      "id": "1909702",
      "postDate": "08/22/2022 20:56:58",
      "content": "<p>Yes defintely! Thanks for so quick a response!</p>",
      "rawMarkdown": "Yes defintely! Thanks for so quick a response!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1909697,
      "author_name": "peterholderrieth",
      "author_url": "",
      "post_date": "08/22/2022 20:55:31",
      "content": "<p>Hi <a href=\"https://www.kaggle.com/comp540cl123\" target=\"_blank\">@comp540cl123</a>! Good catch! This is intended: we derive cell type from RNA data. Therefore, to avoid any leak of information, we do not release RNA for the data set where RNA is the target for the prediction and therefore hidden  (as the case for the Multiome). For the CITEseq data, the RNA for the test is released - as it serves as the input for the predictions for surface protein expressions. Therefore, we released it there. Does that answer your question?</p>",
      "votes": null,
      "replies": [
        {
          "id": 1909702,
          "author_name": "comp540cl123",
          "author_url": "",
          "post_date": "08/22/2022 20:56:58",
          "content": "<p>Yes defintely! Thanks for so quick a response!</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1909624": "Hello there,\n\nI was doing some EDA today with the metadata.csv.\n\nWhile trying to summarize the cell type information, I found that the Cite-seq test data label is available while 10XMultiome ones were hidden. Is it supposed to be so or by mistake?\n\nHere is the plot:\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4455340%2F99638aac0d2205b27ee5e41648053430%2FPicture1.png?generation=1661194992446561&alt=media)",
    "1909697": "Hi @comp540cl123! Good catch! This is intended: we derive cell type from RNA data. Therefore, to avoid any leak of information, we do not release RNA for the data set where RNA is the target for the prediction and therefore hidden  (as the case for the Multiome). For the CITEseq data, the RNA for the test is released - as it serves as the input for the predictions for surface protein expressions. Therefore, we released it there. Does that answer your question?",
    "1909702": "Yes defintely! Thanks for so quick a response!"
  },
  "source": "meta"
}