{ "id": 425031, "title": "External Data", "url": "/competitions/cafa-5-protein-function-prediction/discussion/425031", "author_name": "", "post_date": "2023-07-16T19:21:37.834446800Z", "votes": 1, "comment_count": 1, "views": 0, "content": "

Hello,

\n

Because of the unven distribution of data across different aspects ( BPO, MFO, CCO) I decided to investigate if I could add new data.

\n

I first collected evidence codes which were considered acceptable in the uniprot 2205 release. I came up with the following:
\n[ECO:0000247, ECO:0000250, ECO:0000255, ECO:0000266, ECO:0000269, ECO:0000270, ECO:0000303, ECO:0000304, ECO:0000305, ECO:0000314, ECO:0000315, ECO:0000316, ECO:0000317, ECO:0000318, ECO:0000353, ECO:0007001, ECO:0007003, ECO:0007005, ECO:0007007, ECO:0007669]

\n

I then collected new proteins from the 2302 uniprot release. This gives me 0.5 million additional proteins, 2.9 million immediate go terms and 30 million go terms after propagation. This is a huge amount of data that can make a significant difference to any model.

\n

I am guessing that my list of evidence codes is too broad but this was derived experimentally from the 202205 data. Can the organizers please confirm or release the list of evidence codes that they used to generate the original data for this competition?

\n

All of this is just from https://ftp.uniprot.org/pub/databases/uniprot/current_release/knowledgebase/complete/uniprot_sprot.xml.gz. I have not looked at trembl yet.

\n

Thanks

", "messages": [ { "id": "2347159", "postDate": "07/16/2023 19:21:37", "content": "

Hello,

\n

Because of the unven distribution of data across different aspects ( BPO, MFO, CCO) I decided to investigate if I could add new data.

\n

I first collected evidence codes which were considered acceptable in the uniprot 2205 release. I came up with the following:
\n[ECO:0000247, ECO:0000250, ECO:0000255, ECO:0000266, ECO:0000269, ECO:0000270, ECO:0000303, ECO:0000304, ECO:0000305, ECO:0000314, ECO:0000315, ECO:0000316, ECO:0000317, ECO:0000318, ECO:0000353, ECO:0007001, ECO:0007003, ECO:0007005, ECO:0007007, ECO:0007669]

\n

I then collected new proteins from the 2302 uniprot release. This gives me 0.5 million additional proteins, 2.9 million immediate go terms and 30 million go terms after propagation. This is a huge amount of data that can make a significant difference to any model.

\n

I am guessing that my list of evidence codes is too broad but this was derived experimentally from the 202205 data. Can the organizers please confirm or release the list of evidence codes that they used to generate the original data for this competition?

\n

All of this is just from https://ftp.uniprot.org/pub/databases/uniprot/current_release/knowledgebase/complete/uniprot_sprot.xml.gz. I have not looked at trembl yet.

\n

Thanks

", "rawMarkdown": "Hello,\n\nBecause of the unven distribution of data across different aspects ( BPO, MFO, CCO) I decided to investigate if I could add new data.\n\nI first collected evidence codes which were considered acceptable in the uniprot 2205 release. I came up with the following:\n[ECO:0000247, ECO:0000250, ECO:0000255, ECO:0000266, ECO:0000269, ECO:0000270, ECO:0000303, ECO:0000304, ECO:0000305, ECO:0000314, ECO:0000315, ECO:0000316, ECO:0000317, ECO:0000318, ECO:0000353, ECO:0007001, ECO:0007003, ECO:0007005, ECO:0007007, ECO:0007669]\n\nI then collected new proteins from the 2302 uniprot release. This gives me 0.5 million additional proteins, 2.9 million immediate go terms and 30 million go terms after propagation. This is a huge amount of data that can make a significant difference to any model.\n\nI am guessing that my list of evidence codes is too broad but this was derived experimentally from the 202205 data. Can the organizers please confirm or release the list of evidence codes that they used to generate the original data for this competition?\n\nAll of this is just from https://ftp.uniprot.org/pub/databases/uniprot/current_release/knowledgebase/complete/uniprot_sprot.xml.gz. I have not looked at trembl yet.\n\nThanks", "votes": null }, { "id": "2363422", "postDate": "07/28/2023 16:41:05", "content": "

You can checkout the dataset description page here

\n
\n

Training Set
\n For the training set, we include all proteins with annotated terms that have been validated by experimental or high-throughput evidence, traceable author statement (evidence code TAS), or inferred by curator (IC). More information about evidence codes can be found here. We use annotations from the UniProtKB release of 2022-11-17. The participants are not required to use these data and are also welcome to use any other data available to them.

\n
\n

More details from the File Description section ->

\n
\n

The proteins were all retrieved from the UniProt data set curated at the European Bioinformatics Institute.
\n The header contains the protein's UniProt accession ID and additional information about the protein. Most protein sequences were extracted from the Swiss-Prot database, but a subset of proteins that are not represented in Swiss-Prot were extracted from the TrEMBL database. In both cases, the sequences come from the 2022_05 release from 14-Dec-2022. More information can be found here.

\n

Any sequences taken from TrEMBL will have tr in the header instead of sp. Swiss-Prot and TrEMBL are both parts of UniProtKB.

\n

To obtain the full set of protein sequences for unlabeled proteins, the Swiss-Prot and TrEMBL databases can be found here.

\n
", "rawMarkdown": "You can checkout the dataset description page [here](https://www.kaggle.com/competitions/cafa-5-protein-function-prediction/data)\n>Training Set\nFor the training set, we include all proteins with annotated terms that have been validated by experimental or high-throughput evidence, traceable author statement (evidence code TAS), or inferred by curator (IC). More information about evidence codes can be found here. We use annotations from the UniProtKB release of 2022-11-17. The participants are not required to use these data and are also welcome to use any other data available to them. \n\n\nMore details from the File Description section ->\n \n>The proteins were all retrieved from the UniProt data set curated at the European Bioinformatics Institute.\nThe header contains the protein's UniProt accession ID and additional information about the protein. Most protein sequences were extracted from the Swiss-Prot database, but a subset of proteins that are not represented in Swiss-Prot were extracted from the TrEMBL database. In both cases, the sequences come from the 2022_05 release from 14-Dec-2022. More information can be found here. \n\n \n>Any sequences taken from TrEMBL will have tr in the header instead of sp. Swiss-Prot and TrEMBL are both parts of UniProtKB. \n\n \n>To obtain the full set of protein sequences for unlabeled proteins, the Swiss-Prot and TrEMBL databases can be found [here](https://www.uniprot.org/help/downloads).", "votes": null } ], "comments": [ { "id": 2363422, "author_name": "priyanagda", "author_url": "", "post_date": "07/28/2023 16:41:05", "content": "

You can checkout the dataset description page here

\n
\n

Training Set
\n For the training set, we include all proteins with annotated terms that have been validated by experimental or high-throughput evidence, traceable author statement (evidence code TAS), or inferred by curator (IC). More information about evidence codes can be found here. We use annotations from the UniProtKB release of 2022-11-17. The participants are not required to use these data and are also welcome to use any other data available to them.

\n
\n

More details from the File Description section ->

\n
\n

The proteins were all retrieved from the UniProt data set curated at the European Bioinformatics Institute.
\n The header contains the protein's UniProt accession ID and additional information about the protein. Most protein sequences were extracted from the Swiss-Prot database, but a subset of proteins that are not represented in Swiss-Prot were extracted from the TrEMBL database. In both cases, the sequences come from the 2022_05 release from 14-Dec-2022. More information can be found here.

\n

Any sequences taken from TrEMBL will have tr in the header instead of sp. Swiss-Prot and TrEMBL are both parts of UniProtKB.

\n

To obtain the full set of protein sequences for unlabeled proteins, the Swiss-Prot and TrEMBL databases can be found here.

\n
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "2347159": "Hello,\n\nBecause of the unven distribution of data across different aspects ( BPO, MFO, CCO) I decided to investigate if I could add new data.\n\nI first collected evidence codes which were considered acceptable in the uniprot 2205 release. I came up with the following:\n[ECO:0000247, ECO:0000250, ECO:0000255, ECO:0000266, ECO:0000269, ECO:0000270, ECO:0000303, ECO:0000304, ECO:0000305, ECO:0000314, ECO:0000315, ECO:0000316, ECO:0000317, ECO:0000318, ECO:0000353, ECO:0007001, ECO:0007003, ECO:0007005, ECO:0007007, ECO:0007669]\n\nI then collected new proteins from the 2302 uniprot release. This gives me 0.5 million additional proteins, 2.9 million immediate go terms and 30 million go terms after propagation. This is a huge amount of data that can make a significant difference to any model.\n\nI am guessing that my list of evidence codes is too broad but this was derived experimentally from the 202205 data. Can the organizers please confirm or release the list of evidence codes that they used to generate the original data for this competition?\n\nAll of this is just from https://ftp.uniprot.org/pub/databases/uniprot/current_release/knowledgebase/complete/uniprot_sprot.xml.gz. I have not looked at trembl yet.\n\nThanks", "2363422": "You can checkout the dataset description page [here](https://www.kaggle.com/competitions/cafa-5-protein-function-prediction/data)\n>Training Set\nFor the training set, we include all proteins with annotated terms that have been validated by experimental or high-throughput evidence, traceable author statement (evidence code TAS), or inferred by curator (IC). More information about evidence codes can be found here. We use annotations from the UniProtKB release of 2022-11-17. The participants are not required to use these data and are also welcome to use any other data available to them. \n\n\nMore details from the File Description section ->\n \n>The proteins were all retrieved from the UniProt data set curated at the European Bioinformatics Institute.\nThe header contains the protein's UniProt accession ID and additional information about the protein. Most protein sequences were extracted from the Swiss-Prot database, but a subset of proteins that are not represented in Swiss-Prot were extracted from the TrEMBL database. In both cases, the sequences come from the 2022_05 release from 14-Dec-2022. More information can be found here. \n\n \n>Any sequences taken from TrEMBL will have tr in the header instead of sp. Swiss-Prot and TrEMBL are both parts of UniProtKB. \n\n \n>To obtain the full set of protein sequences for unlabeled proteins, the Swiss-Prot and TrEMBL databases can be found [here](https://www.uniprot.org/help/downloads)." }, "source": "meta" }