Hi all, we wanted to share updates on an additional resource we have uploaded to the competition dataset. We have been preparing MSAs for the train_v2 dataset and have uploaded the subset that have completed (2534 files) to MSA_v2/. These files were generated from the outputs of rMSA and are in the A2M fasta format.
\nPlease see more details on how we generated these files in our previous forum post (https://www.kaggle.com/competitions/stanford-rna-3d-folding/discussion/569424).
\nAdditionally, to note, there are some target ids in the train_v2 dataset that are case insensitive duplicates (e.g. 4W2E_A and 4W2E_a). Kaggle does not allow file names that differ only in case and, as there are only 37 of these pairs, we will exclude the duplicates from the MSA upload.
Please note that at test time, your Notebook should still look in MSA/ for MSA files for test sequences. This new folder MSA_v2/ is for competitors who are training on the expanded train_v2 dataset.
", "messages": [ { "id": 3199899, "postDate": "2025-05-11T16:42:27.597Z", "content": "Hi all, we wanted to share updates on an additional resource we have uploaded to the competition dataset. We have been preparing MSAs for the train_v2 dataset and have uploaded the subset that have completed (2534 files) to MSA_v2/. These files were generated from the outputs of rMSA and are in the A2M fasta format.
\nPlease see more details on how we generated these files in our previous forum post (https://www.kaggle.com/competitions/stanford-rna-3d-folding/discussion/569424).
\nAdditionally, to note, there are some target ids in the train_v2 dataset that are case insensitive duplicates (e.g. 4W2E_A and 4W2E_a). Kaggle does not allow file names that differ only in case and, as there are only 37 of these pairs, we will exclude the duplicates from the MSA upload.
Please note that at test time, your Notebook should still look in MSA/ for MSA files for test sequences. This new folder MSA_v2/ is for competitors who are training on the expanded train_v2 dataset.
", "rawMarkdown": "Hi all, we wanted to share updates on an additional resource we have uploaded to the competition dataset. We have been preparing MSAs for the train_v2 dataset and have uploaded the subset that have completed (2534 files) to **MSA_v2/**. These files were generated from the outputs of rMSA and are in the A2M fasta format. \n\nPlease see more details on how we generated these files in our previous forum post (https://www.kaggle.com/competitions/stanford-rna-3d-folding/discussion/569424).\n\nAdditionally, to note, there are some target ids in the train_v2 dataset that are case insensitive duplicates (e.g. `4W2E_A` and `4W2E_a`). Kaggle does not allow file names that differ only in case and, as there are only 37 of these pairs, we will exclude the duplicates from the MSA upload.\n\nPlease note that at test time, your Notebook should still look in **MSA/** for MSA files for test sequences. This new folder **MSA_v2/** is for competitors who are training on the expanded train_v2 dataset.\n\n\n\n", "votes": 8 }, { "id": 3201039, "postDate": "2025-05-13T11:20:10.013Z", "content": "@alissahummer , is it possible that discussions started by competition hosts are pinned so that they can be found at the beginning of discussions? Thank you very much.
", "rawMarkdown": " @alissahummer , is it possible that discussions started by competition hosts are pinned so that they can be found at the beginning of discussions? Thank you very much." }, { "id": 3201410, "postDate": "2025-05-13T20:24:20.563Z", "rawMarkdown": "", "isDeleted": true } ], "comments": [ { "id": 3201039, "author_name": "gmobaz", "author_url": "", "post_date": "2025-05-13T11:20:10.013000", "content": "@alissahummer , is it possible that discussions started by competition hosts are pinned so that they can be found at the beginning of discussions? Thank you very much.
", "votes": 0, "replies": [] }, { "id": 3201410, "author_name": "", "author_url": "", "post_date": "2025-05-13T20:24:20.563000", "content": "", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "3199899": "Hi all, we wanted to share updates on an additional resource we have uploaded to the competition dataset. We have been preparing MSAs for the train_v2 dataset and have uploaded the subset that have completed (2534 files) to **MSA_v2/**. These files were generated from the outputs of rMSA and are in the A2M fasta format. \n\nPlease see more details on how we generated these files in our previous forum post (https://www.kaggle.com/competitions/stanford-rna-3d-folding/discussion/569424).\n\nAdditionally, to note, there are some target ids in the train_v2 dataset that are case insensitive duplicates (e.g. `4W2E_A` and `4W2E_a`). Kaggle does not allow file names that differ only in case and, as there are only 37 of these pairs, we will exclude the duplicates from the MSA upload.\n\nPlease note that at test time, your Notebook should still look in **MSA/** for MSA files for test sequences. This new folder **MSA_v2/** is for competitors who are training on the expanded train_v2 dataset.\n\n\n\n", "3201039": " @alissahummer , is it possible that discussions started by competition hosts are pinned so that they can be found at the beginning of discussions? Thank you very much.", "3201410": "" } }