{
  "id": 417561,
  "title": "SPROF-GO predictions",
  "url": "/competitions/cafa-5-protein-function-prediction/discussion/417561",
  "author_name": "",
  "post_date": "2023-06-16T08:28:00.739841900Z",
  "votes": 19,
  "comment_count": 5,
  "views": 0,
  "content": "<p>SPROF-GO (<a href=\"https://github.com/biomed-AI/SPROF-GO\" target=\"_blank\">https://github.com/biomed-AI/SPROF-GO</a>) is a recently published method for protein function prediction. <br>\nI run the program on colab and here I'm sharing the output for the test set (score 0.50148): </p>\n<p><a href=\"https://www.kaggle.com/code/mtinti/sprof-predictions/\" target=\"_blank\">https://www.kaggle.com/code/mtinti/sprof-predictions/</a></p>\n<p>For the predictions, I truncated the proteins greater than 2000 aa.</p>",
  "messages": [
    {
      "id": "2304780",
      "postDate": "06/16/2023 08:28:00",
      "content": "<p>SPROF-GO (<a href=\"https://github.com/biomed-AI/SPROF-GO\" target=\"_blank\">https://github.com/biomed-AI/SPROF-GO</a>) is a recently published method for protein function prediction. <br>\nI run the program on colab and here I'm sharing the output for the test set (score 0.50148): </p>\n<p><a href=\"https://www.kaggle.com/code/mtinti/sprof-predictions/\" target=\"_blank\">https://www.kaggle.com/code/mtinti/sprof-predictions/</a></p>\n<p>For the predictions, I truncated the proteins greater than 2000 aa.</p>",
      "rawMarkdown": "SPROF-GO (https://github.com/biomed-AI/SPROF-GO) is a recently published method for protein function prediction. \nI run the program on colab and here I'm sharing the output for the test set (score 0.50148): \n\nhttps://www.kaggle.com/code/mtinti/sprof-predictions/\n\nFor the predictions, I truncated the proteins greater than 2000 aa.",
      "votes": null
    },
    {
      "id": "2305557",
      "postDate": "06/16/2023 19:03:03",
      "content": "<p>Did you retrain the model using cafa5's training data? Thanks!</p>",
      "rawMarkdown": "Did you retrain the model using cafa5's training data? Thanks!",
      "votes": null
    },
    {
      "id": "2305625",
      "postDate": "06/16/2023 20:01:30",
      "content": "<p>No, I didn't look into it as it didn't seem strightforward.  I just made predictions. </p>",
      "rawMarkdown": "No, I didn't look into it as it didn't seem strightforward.  I just made predictions.",
      "votes": null
    },
    {
      "id": "2305981",
      "postDate": "06/17/2023 03:39:29",
      "content": "<p>Thank you for sharing! Is this method significantly improved by using sequence embedding (shapes like (b, L, feature_dim))？ Because this requires larger computing resources, I haven't tried it yet.</p>",
      "rawMarkdown": "Thank you for sharing! Is this method significantly improved by using sequence embedding (shapes like (b, L, feature_dim))？ Because this requires larger computing resources, I haven't tried it yet.",
      "votes": null
    },
    {
      "id": "2306221",
      "postDate": "06/17/2023 06:46:49",
      "content": "<p>hi, I'm not sure I understand your first question. However, I was just curious to have a sort of base score from a fairly recent method.</p>",
      "rawMarkdown": "hi, I'm not sure I understand your first question. However, I was just curious to have a sort of base score from a fairly recent method.",
      "votes": null
    },
    {
      "id": "3346183",
      "postDate": "11/24/2025 06:53:19",
      "content": "<p>Hi, what is the top k did you set in the prediction</p>",
      "rawMarkdown": "Hi, what is the top k did you set in the prediction",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2305557,
      "author_name": "evansliu",
      "author_url": "",
      "post_date": "06/16/2023 19:03:03",
      "content": "<p>Did you retrain the model using cafa5's training data? Thanks!</p>",
      "votes": null,
      "replies": [
        {
          "id": 2305625,
          "author_name": "mtinti",
          "author_url": "",
          "post_date": "06/16/2023 20:01:30",
          "content": "<p>No, I didn't look into it as it didn't seem strightforward.  I just made predictions. </p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 2305981,
      "author_name": "xiedonglin",
      "author_url": "",
      "post_date": "06/17/2023 03:39:29",
      "content": "<p>Thank you for sharing! Is this method significantly improved by using sequence embedding (shapes like (b, L, feature_dim))？ Because this requires larger computing resources, I haven't tried it yet.</p>",
      "votes": null,
      "replies": [
        {
          "id": 2306221,
          "author_name": "mtinti",
          "author_url": "",
          "post_date": "06/17/2023 06:46:49",
          "content": "<p>hi, I'm not sure I understand your first question. However, I was just curious to have a sort of base score from a fairly recent method.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 3346183,
      "author_name": "aizealotwu",
      "author_url": "",
      "post_date": "11/24/2025 06:53:19",
      "content": "<p>Hi, what is the top k did you set in the prediction</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2304780": "SPROF-GO (https://github.com/biomed-AI/SPROF-GO) is a recently published method for protein function prediction. \nI run the program on colab and here I'm sharing the output for the test set (score 0.50148): \n\nhttps://www.kaggle.com/code/mtinti/sprof-predictions/\n\nFor the predictions, I truncated the proteins greater than 2000 aa.",
    "2305557": "Did you retrain the model using cafa5's training data? Thanks!",
    "2305625": "No, I didn't look into it as it didn't seem strightforward.  I just made predictions.",
    "2305981": "Thank you for sharing! Is this method significantly improved by using sequence embedding (shapes like (b, L, feature_dim))？ Because this requires larger computing resources, I haven't tried it yet.",
    "2306221": "hi, I'm not sure I understand your first question. However, I was just curious to have a sort of base score from a fairly recent method.",
    "3346183": "Hi, what is the top k did you set in the prediction"
  },
  "source": "meta"
}