Hello community! Would be very grateful for any advice. As mentioned in the description, the location column in the multiome_train.parquet file contains gene symbols. But when I filter only genes present in the de_train file (column names), it appears that there are only 15580 of the 18211 genes.
\nDo I understand correctly that we only have multiome data (baseline expression) for a subset of genes, or am I doing something wrong?
", "messages": [ { "id": "2462242", "postDate": "09/30/2023 08:18:35", "content": "Hello community! Would be very grateful for any advice. As mentioned in the description, the location column in the multiome_train.parquet file contains gene symbols. But when I filter only genes present in the de_train file (column names), it appears that there are only 15580 of the 18211 genes.
\nDo I understand correctly that we only have multiome data (baseline expression) for a subset of genes, or am I doing something wrong?
", "rawMarkdown": "Hello community! Would be very grateful for any advice. As mentioned in the description, the location column in the multiome_train.parquet file contains gene symbols. But when I filter only genes present in the de_train file (column names), it appears that there are only 15580 of the 18211 genes. \n\nDo I understand correctly that we only have multiome data (baseline expression) for a subset of genes, or am I doing something wrong?", "votes": null }, { "id": "2463784", "postDate": "10/01/2023 14:57:53", "content": "Maybe we both doing something wrong - I see the same number of missing genes.
", "rawMarkdown": "Maybe we both doing something wrong - I see the same number of missing genes.", "votes": null }, { "id": "2463850", "postDate": "10/01/2023 16:07:19", "content": "there was a post about this a while ago, here: https://www.kaggle.com/competitions/open-problems-single-cell-perturbations/discussion/440792
", "rawMarkdown": "there was a post about this a while ago, here: https://www.kaggle.com/competitions/open-problems-single-cell-perturbations/discussion/440792", "votes": null }, { "id": "2464089", "postDate": "10/01/2023 23:52:59", "content": "Think this is a different question. There are 2631 genes which have no multiome_var_meta data. Makes for a pretty big hole if you wanted to add 'chromosome' as a feature in your model.
", "rawMarkdown": "Think this is a different question. There are 2631 genes which have no multiome_var_meta data. Makes for a pretty big hole if you wanted to add 'chromosome' as a feature in your model.", "votes": null }, { "id": "2464115", "postDate": "10/02/2023 00:30:22", "content": "yeah, I saw it. It's about adata_train.parquet, says it is actually a superset of de_train. Here is a different situation - multiome data has **fewer **genes than de_train, unless I did not understand smth…
", "rawMarkdown": "yeah, I saw it. It's about adata_train.parquet, says it is actually a superset of de_train. Here is a different situation - multiome data has **fewer **genes than de_train, unless I did not understand smth...", "votes": null } ], "comments": [ { "id": 2463784, "author_name": "pcjimmmy", "author_url": "", "post_date": "10/01/2023 14:57:53", "content": "Maybe we both doing something wrong - I see the same number of missing genes.
", "votes": null, "replies": [] }, { "id": 2463850, "author_name": "rahi37", "author_url": "", "post_date": "10/01/2023 16:07:19", "content": "there was a post about this a while ago, here: https://www.kaggle.com/competitions/open-problems-single-cell-perturbations/discussion/440792
", "votes": null, "replies": [ { "id": 2464089, "author_name": "pcjimmmy", "author_url": "", "post_date": "10/01/2023 23:52:59", "content": "Think this is a different question. There are 2631 genes which have no multiome_var_meta data. Makes for a pretty big hole if you wanted to add 'chromosome' as a feature in your model.
", "votes": null, "replies": [] }, { "id": 2464115, "author_name": "antoninadolgorukova", "author_url": "", "post_date": "10/02/2023 00:30:22", "content": "yeah, I saw it. It's about adata_train.parquet, says it is actually a superset of de_train. Here is a different situation - multiome data has **fewer **genes than de_train, unless I did not understand smth…
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "2462242": "Hello community! Would be very grateful for any advice. As mentioned in the description, the location column in the multiome_train.parquet file contains gene symbols. But when I filter only genes present in the de_train file (column names), it appears that there are only 15580 of the 18211 genes. \n\nDo I understand correctly that we only have multiome data (baseline expression) for a subset of genes, or am I doing something wrong?", "2463784": "Maybe we both doing something wrong - I see the same number of missing genes.", "2463850": "there was a post about this a while ago, here: https://www.kaggle.com/competitions/open-problems-single-cell-perturbations/discussion/440792", "2464089": "Think this is a different question. There are 2631 genes which have no multiome_var_meta data. Makes for a pretty big hole if you wanted to add 'chromosome' as a feature in your model.", "2464115": "yeah, I saw it. It's about adata_train.parquet, says it is actually a superset of de_train. Here is a different situation - multiome data has **fewer **genes than de_train, unless I did not understand smth..." }, "source": "meta" }