After reading a recent paper in which it was applied to the prediction of antibody interface residues, I wonder if it would be possible to get meaningful features summarizing the molecule structures by 3D Zernike descriptors. Has anyone tried something similar already? I might give a try later if I find the time.
", "messages": [ { "id": "547369", "postDate": "06/07/2019 16:18:33", "content": "After reading a recent paper in which it was applied to the prediction of antibody interface residues, I wonder if it would be possible to get meaningful features summarizing the molecule structures by 3D Zernike descriptors. Has anyone tried something similar already? I might give a try later if I find the time.
", "rawMarkdown": "After reading a [recent paper in which it was applied to the prediction of antibody interface residues](https://academic.oup.com/bioinformatics/article-abstract/35/11/1870/5161081?redirectedFrom=fulltext), I wonder if it would be possible to get meaningful features summarizing the molecule structures by [3D Zernike descriptors](http://cg.cs.uni-bonn.de/aigaion2root/attachments/novotni-2003-3d.pdf). Has anyone tried something similar already? I might give a try later if I find the time.", "votes": null }, { "id": "547475", "postDate": "06/07/2019 19:15:50", "content": "Sounds like a pretty complicated predictor. You may want to try with easier ones first!
", "rawMarkdown": "Sounds like a pretty complicated predictor. You may want to try with easier ones first!", "votes": null }, { "id": "547486", "postDate": "06/07/2019 19:40:55", "content": "I thought about using them as they are one of the most commonly used rotation-invariant descriptors but haven't really had the time to do much with this competition. There are also spherical harmonics which could be very useful in this context. I don't think they will make good predictors by themselves, but may be useful depending on how you use them exactly.
", "rawMarkdown": "I thought about using them as they are one of the most commonly used rotation-invariant descriptors but haven't really had the time to do much with this competition. There are also spherical harmonics which could be very useful in this context. I don't think they will make good predictors by themselves, but may be useful depending on how you use them exactly.", "votes": null } ], "comments": [ { "id": 547475, "author_name": "borisdee", "author_url": "", "post_date": "06/07/2019 19:15:50", "content": "Sounds like a pretty complicated predictor. You may want to try with easier ones first!
", "votes": null, "replies": [] }, { "id": 547486, "author_name": "matthewmasters", "author_url": "", "post_date": "06/07/2019 19:40:55", "content": "I thought about using them as they are one of the most commonly used rotation-invariant descriptors but haven't really had the time to do much with this competition. There are also spherical harmonics which could be very useful in this context. I don't think they will make good predictors by themselves, but may be useful depending on how you use them exactly.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "547369": "After reading a [recent paper in which it was applied to the prediction of antibody interface residues](https://academic.oup.com/bioinformatics/article-abstract/35/11/1870/5161081?redirectedFrom=fulltext), I wonder if it would be possible to get meaningful features summarizing the molecule structures by [3D Zernike descriptors](http://cg.cs.uni-bonn.de/aigaion2root/attachments/novotni-2003-3d.pdf). Has anyone tried something similar already? I might give a try later if I find the time.", "547475": "Sounds like a pretty complicated predictor. You may want to try with easier ones first!", "547486": "I thought about using them as they are one of the most commonly used rotation-invariant descriptors but haven't really had the time to do much with this competition. There are also spherical harmonics which could be very useful in this context. I don't think they will make good predictors by themselves, but may be useful depending on how you use them exactly." }, "source": "meta" }