Hey everyone, while researching for additional realworld datasets related to cryoET I have discovered following datasets which I think would be useful for pretraining the models:
\n\nCryo-electron microscopy (cryo-EM) a powerful technique determining the structures biological macromolecular complexes. Picking single-protein particles cryo-EM micrographs a crucial step reconstructing protein structures. However, the widely used -based particle picking process labor-intensive -consuming. Though machine learning artificial intelligence (AI) based particle picking can potentially automate the process, its development hindered lack , high-quality labelled training data. address this bottleneck, we present CryoPPP, a , diverse, expert-curated cryo-EM image dataset protein particle picking analysis. It consists labelled cryo-EM micrographs (images) representative protein datasets selected the Electron Microscopy Image Archive (EMPIAR). The dataset terabytes includes , high-resolution micrographs labelled protein particle coordinates. The labelling process was rigorously validated through D particle validation D density map validation the gold standard. The dataset expected greatly facilitate the development AI classical methods automated cryo-EM protein particle picking.\n2 Here are some of the repo using this dataset:
\ni. https://github.com/jianlin-cheng/CryoTransformer
\nii. https://github.com/jianlin-cheng/CryoSegNet
\nI am currently investigating on how to incorporate it in my workflow, let me know if any you have some ideas regarding this!
\nAdditional Note: The above dataset contains same particles like beta-galactosidase, ribosomes and VLPs.
Hey everyone, while researching for additional realworld datasets related to cryoET I have discovered following datasets which I think would be useful for pretraining the models:
\n\nCryo-electron microscopy (cryo-EM) a powerful technique determining the structures biological macromolecular complexes. Picking single-protein particles cryo-EM micrographs a crucial step reconstructing protein structures. However, the widely used -based particle picking process labor-intensive -consuming. Though machine learning artificial intelligence (AI) based particle picking can potentially automate the process, its development hindered lack , high-quality labelled training data. address this bottleneck, we present CryoPPP, a , diverse, expert-curated cryo-EM image dataset protein particle picking analysis. It consists labelled cryo-EM micrographs (images) representative protein datasets selected the Electron Microscopy Image Archive (EMPIAR). The dataset terabytes includes , high-resolution micrographs labelled protein particle coordinates. The labelling process was rigorously validated through D particle validation D density map validation the gold standard. The dataset expected greatly facilitate the development AI classical methods automated cryo-EM protein particle picking.\n2 Here are some of the repo using this dataset:
\ni. https://github.com/jianlin-cheng/CryoTransformer
\nii. https://github.com/jianlin-cheng/CryoSegNet
\nI am currently investigating on how to incorporate it in my workflow, let me know if any you have some ideas regarding this!
\nAdditional Note: The above dataset contains same particles like beta-galactosidase, ribosomes and VLPs.
Good job! THanks
", "rawMarkdown": "Good job! THanks", "votes": null }, { "id": "3082077", "postDate": "12/27/2024 16:22:50", "content": "Thanks for share. Is more tha nothing but according to description apo-ferritin and thyroglobulin are missing. Asuming \"Bacteriophage MS2\" is virus.
", "rawMarkdown": "Thanks for share. Is more tha nothing but according to description apo-ferritin and thyroglobulin are missing. Asuming \"Bacteriophage MS2\" is virus.", "votes": null }, { "id": "3083768", "postDate": "12/30/2024 02:17:50", "content": "Thanks for sharing! I noticed that these datasets are derived from cryoEM data, but our task involves cryoET, and it’s clear that they differ in the preprocessing stages. I think it might be a good idea to explore some self-supervised methods to pretrain on cryoEM data (as cryoEM data is likely to be far more abundant than cryoET data) and then fine-tune on cryoET data in the second stage. Finally, transferring the model to specific tasks (such as our competition) could potentially
", "rawMarkdown": "Thanks for sharing! I noticed that these datasets are derived from cryoEM data, but our task involves cryoET, and it’s clear that they differ in the preprocessing stages. I think it might be a good idea to explore some self-supervised methods to pretrain on cryoEM data (as cryoEM data is likely to be far more abundant than cryoET data) and then fine-tune on cryoET data in the second stage. Finally, transferring the model to specific tasks (such as our competition) could potentially", "votes": null } ], "comments": [ { "id": 3081843, "author_name": "mrsimple07", "author_url": "", "post_date": "12/27/2024 09:29:21", "content": "Good job! THanks
", "votes": null, "replies": [] }, { "id": 3082077, "author_name": "sacuscreed", "author_url": "", "post_date": "12/27/2024 16:22:50", "content": "Thanks for share. Is more tha nothing but according to description apo-ferritin and thyroglobulin are missing. Asuming \"Bacteriophage MS2\" is virus.
", "votes": null, "replies": [] }, { "id": 3083768, "author_name": "lonelycrab", "author_url": "", "post_date": "12/30/2024 02:17:50", "content": "Thanks for sharing! I noticed that these datasets are derived from cryoEM data, but our task involves cryoET, and it’s clear that they differ in the preprocessing stages. I think it might be a good idea to explore some self-supervised methods to pretrain on cryoEM data (as cryoEM data is likely to be far more abundant than cryoET data) and then fine-tune on cryoET data in the second stage. Finally, transferring the model to specific tasks (such as our competition) could potentially
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "3081774": "Hey everyone, while researching for additional realworld datasets related to cryoET I have discovered following datasets which I think would be useful for pretraining the models:\n\n1. https://github.com/BioinfoMachineLearning/cryoppp\n\n```\nCryo-electron microscopy (cryo-EM) is a powerful technique for determining the structures of biological macromolecular complexes. Picking single-protein particles from cryo-EM micrographs is a crucial step in reconstructing protein structures. However, the widely used template-based particle picking process is labor-intensive and time-consuming. Though machine learning and artificial intelligence (AI) based particle picking can potentially automate the process, its development is hindered by lack of large, high-quality labelled training data. To address this bottleneck, we present CryoPPP, a large, diverse, expert-curated cryo-EM image dataset for protein particle picking and analysis. It consists of labelled cryo-EM micrographs (images) of 34 representative protein datasets selected from the Electron Microscopy Public Image Archive (EMPIAR). The dataset is 2.6 terabytes and includes 9,893 high-resolution micrographs with labelled protein particle coordinates. The labelling process was rigorously validated through 2D particle class validation and 3D density map validation with the gold standard. The dataset is expected to greatly facilitate the development of both AI and classical methods for automated cryo-EM protein particle picking.\n```\n2 Here are some of the repo using this dataset:\n\ni. https://github.com/jianlin-cheng/CryoTransformer\nii. https://github.com/jianlin-cheng/CryoSegNet\n\n\nI am currently investigating on how to incorporate it in my workflow, let me know if any you have some ideas regarding this!\n\n\nAdditional Note: The above dataset contains same particles like beta-galactosidase, ribosomes and VLPs.", "3081843": "Good job! THanks", "3082077": "Thanks for share. Is more tha nothing but according to description apo-ferritin and thyroglobulin are missing. Asuming \"Bacteriophage MS2\" is virus.", "3083768": "Thanks for sharing! I noticed that these datasets are derived from cryoEM data, but our task involves cryoET, and it’s clear that they differ in the preprocessing stages. I think it might be a good idea to explore some self-supervised methods to pretrain on cryoEM data (as cryoEM data is likely to be far more abundant than cryoET data) and then fine-tune on cryoET data in the second stage. Finally, transferring the model to specific tasks (such as our competition) could potentially" }, "source": "meta" }