Can someone brief me the biological part of the problem. How does reactivity relates to the structure of the RNA? Isn't the RNA sequence enough to determine the structure using dynamic programming? Do we have to predict the reactivity which are null valued in the training set?
", "messages": [ { "id": 2434764, "postDate": "2023-09-12T13:54:22.280Z", "content": "Reactivity is the data result of chemical probing experiments where the folded molecules (sequences) are exposed to a reagent (2A3 or DMS) that attacks the molecule. The bases that are paired are protected from exposure and have no reactivity. The bases that are unpaired are exposed and have high reactivity. Naturally, the process and data are not as exact as my description implies.
\nSomething that hasn't been noted elsewhere yet is that 2A3 reacts with all bases whereas DMS only reacts with A and C. See Wikipedia DMS under \"Reactions with nucleic acids\".
\nChemical probing is widely used in RNA structure research due to its lower cost and shorter timeframe compared with the 3D structure determination methods used to generate PDB files. Cryo-EM is particularly challenging with RNA due to its small size.
", "rawMarkdown": "Reactivity is the data result of chemical probing experiments where the folded molecules (sequences) are exposed to a reagent (2A3 or DMS) that attacks the molecule. The bases that are paired are protected from exposure and have no reactivity. The bases that are unpaired are exposed and have high reactivity. Naturally, the process and data are not as exact as my description implies.\n\nSomething that hasn't been noted elsewhere yet is that 2A3 reacts with all bases whereas DMS only reacts with A and C. See [Wikipedia DMS](https://en.wikipedia.org/wiki/Dimethyl_sulfate) under \"Reactions with nucleic acids\".\n\nChemical probing is widely used in RNA structure research due to its lower cost and shorter timeframe compared with the 3D structure determination methods used to generate PDB files. Cryo-EM is particularly challenging with RNA due to its small size.", "votes": 10, "replies": [ { "id": 2435175, "postDate": "2023-09-12T18:40:04.383Z", "content": "Thanks for sharing !
\nAre there some other things \" hasn't been noted elsewhere \" ? )))
Correction: The DMS experiment uses 4-base DMS as discussed HERE. The A and C bases have higher signal to noise on average than the G and U bases in the DMS datasets.
", "rawMarkdown": "Correction: The DMS experiment uses 4-base DMS as discussed [HERE](https://academic.oup.com/nar/article/51/16/8744/7201944). The A and C bases have higher signal to noise on average than the G and U bases in the DMS datasets.", "votes": 5 }, { "id": 2467985, "postDate": "2023-10-05T04:49:43.407Z", "content": "By signal to noise, do you mean dividing each nucleotide's reactivity by its reactivity error?
", "rawMarkdown": "By signal to noise, do you mean dividing each nucleotide's reactivity by its reactivity error?", "votes": 1, "isDeleted": true } ] }, { "id": 2436109, "postDate": "2023-09-13T11:03:30.790Z", "content": "I guess chemical probing was the keyword.
\nhttps://ars.els-cdn.com/content/image/1-s2.0-S0022283622002157-gr1_lrg.jpg
\nHere for all non-bio people who are confused
Thanks man…your answer is concise
", "rawMarkdown": "Thanks man...your answer is concise" } ] }, { "id": 2435006, "postDate": "2023-09-12T16:36:03.197Z", "content": "\n\nIsn't the RNA sequence enough to determine the structure using dynamic programming?
\n
Unfortunately, no! Consider the very short sequence GGAAAACC. This could have the following sets of pairs:
\nConventionally structure is predicted using dynamic programming and comes up with the best option based on measured thermodynamic properties, but those predicted structures still have issues - and conventionally don't take into account pseudoknots.
\n\n\nDo we have to predict the reactivity which are null valued in the training set?
\n
According to the sample submission and information on the home page, it looks like it. I'd assume that when data is unavailable it wouldn't be factored into the scoring, but I'll leave that to be confirmed by someone who has more details on the scoring
", "rawMarkdown": "> Isn't the RNA sequence enough to determine the structure using dynamic programming?\n\nUnfortunately, no! Consider the very short sequence GGAAAACC. This could have the following sets of pairs:\n* 1-8, 2-7\n* 1-8\n* 2-7\n* 1-7\n* 2-8\n* no pairs at all\n\nConventionally structure is predicted using dynamic programming and comes up with the best option based on measured thermodynamic properties, but those predicted structures still have issues - and conventionally don't take into account pseudoknots.\n\n> Do we have to predict the reactivity which are null valued in the training set?\n\nAccording to the sample submission and information on the home page, it looks like it. I'd _assume_ that when data is unavailable it wouldn't be factored into the scoring, but I'll leave that to be confirmed by someone who has more details on the scoring", "votes": 5, "replies": [ { "id": 2439487, "postDate": "2023-09-14T22:54:09.457Z", "content": "How do you know the pairing numbers to those aminoacids? Is there a table that explains it?
", "rawMarkdown": "How do you know the pairing numbers to those aminoacids? Is there a table that explains it?", "votes": 1, "replies": [ { "id": 2439502, "postDate": "2023-09-14T23:40:22.840Z", "content": "\"Canonically\", A pairs with U, and G pairs with C or U. GC being strongest, then AU, then GU. GU is technically a special case but is often considered when determining structures. Also technically you wind up with weaker interactions between other bases (eg, G and A), but are not typically represented in structural models. At any rate what I just listed out was all possible combinations of GC, AU, and GU pairs (in this case, just GC pairs as there are no Us in that sequence)
", "rawMarkdown": "\"Canonically\", A pairs with U, and G pairs with C or U. GC being strongest, then AU, then GU. GU is technically a special case but is often considered when determining structures. Also technically you wind up with weaker interactions between other bases (eg, G and A), but are not typically represented in structural models. At any rate what I just listed out was all possible combinations of GC, AU, and GU pairs (in this case, just GC pairs as there are no Us in that sequence)", "votes": 2 } ] } ] }, { "id": 2434448, "postDate": "2023-09-12T09:50:43.410Z", "content": "Can someone brief me the biological part of the problem. How does reactivity relates to the structure of the RNA? Isn't the RNA sequence enough to determine the structure using dynamic programming? Do we have to predict the reactivity which are null valued in the training set?
", "rawMarkdown": "Can someone brief me the biological part of the problem. How does reactivity relates to the structure of the RNA? Isn't the RNA sequence enough to determine the structure using dynamic programming? Do we have to predict the reactivity which are null valued in the training set?", "votes": 2 }, { "id": 2449555, "postDate": "2023-09-21T08:55:29.107Z", "content": "https://www.kaggle.com/competitions/stanford-ribonanza-rna-folding/discussion/442089 hope it will helpful … Sir @danishmirza0321
", "rawMarkdown": "https://www.kaggle.com/competitions/stanford-ribonanza-rna-folding/discussion/442089 hope it will helpful ... Sir @danishmirza0321 " }, { "id": 2453212, "postDate": "2023-09-23T21:49:11.247Z", "rawMarkdown": "", "isDeleted": true } ], "comments": [ { "id": 2434764, "author_name": "DigitalEmbrace", "author_url": "", "post_date": "2023-09-12T13:54:22.280000", "content": "Reactivity is the data result of chemical probing experiments where the folded molecules (sequences) are exposed to a reagent (2A3 or DMS) that attacks the molecule. The bases that are paired are protected from exposure and have no reactivity. The bases that are unpaired are exposed and have high reactivity. Naturally, the process and data are not as exact as my description implies.
\nSomething that hasn't been noted elsewhere yet is that 2A3 reacts with all bases whereas DMS only reacts with A and C. See Wikipedia DMS under \"Reactions with nucleic acids\".
\nChemical probing is widely used in RNA structure research due to its lower cost and shorter timeframe compared with the 3D structure determination methods used to generate PDB files. Cryo-EM is particularly challenging with RNA due to its small size.
", "votes": 10, "replies": [ { "id": 2435175, "author_name": "Alexander Chervov", "author_url": "", "post_date": "2023-09-12T18:40:04.383000", "content": "Thanks for sharing !
\nAre there some other things \" hasn't been noted elsewhere \" ? )))
Correction: The DMS experiment uses 4-base DMS as discussed HERE. The A and C bases have higher signal to noise on average than the G and U bases in the DMS datasets.
", "votes": 5, "replies": [] }, { "id": 2467985, "author_name": "", "author_url": "", "post_date": "2023-10-05T04:49:43.407000", "content": "By signal to noise, do you mean dividing each nucleotide's reactivity by its reactivity error?
", "votes": 1, "replies": [] } ] }, { "id": 2436109, "author_name": "Danish Mirza", "author_url": "", "post_date": "2023-09-13T11:03:30.790000", "content": "I guess chemical probing was the keyword.
\nhttps://ars.els-cdn.com/content/image/1-s2.0-S0022283622002157-gr1_lrg.jpg
\nHere for all non-bio people who are confused
Thanks man…your answer is concise
", "votes": 0, "replies": [] } ] }, { "id": 2435006, "author_name": "Jonathan Romano", "author_url": "", "post_date": "2023-09-12T16:36:03.197000", "content": "\n\nIsn't the RNA sequence enough to determine the structure using dynamic programming?
\n
Unfortunately, no! Consider the very short sequence GGAAAACC. This could have the following sets of pairs:
\nConventionally structure is predicted using dynamic programming and comes up with the best option based on measured thermodynamic properties, but those predicted structures still have issues - and conventionally don't take into account pseudoknots.
\n\n\nDo we have to predict the reactivity which are null valued in the training set?
\n
According to the sample submission and information on the home page, it looks like it. I'd assume that when data is unavailable it wouldn't be factored into the scoring, but I'll leave that to be confirmed by someone who has more details on the scoring
", "votes": 5, "replies": [ { "id": 2439487, "author_name": "Luisa Frugoli Lopes", "author_url": "", "post_date": "2023-09-14T22:54:09.457000", "content": "How do you know the pairing numbers to those aminoacids? Is there a table that explains it?
", "votes": 1, "replies": [ { "id": 2439502, "author_name": "Jonathan Romano", "author_url": "", "post_date": "2023-09-14T23:40:22.840000", "content": "\"Canonically\", A pairs with U, and G pairs with C or U. GC being strongest, then AU, then GU. GU is technically a special case but is often considered when determining structures. Also technically you wind up with weaker interactions between other bases (eg, G and A), but are not typically represented in structural models. At any rate what I just listed out was all possible combinations of GC, AU, and GU pairs (in this case, just GC pairs as there are no Us in that sequence)
", "votes": 2, "replies": [] } ] } ] }, { "id": 2449555, "author_name": "Malik Muhammad Ahmed", "author_url": "", "post_date": "2023-09-21T08:55:29.107000", "content": "https://www.kaggle.com/competitions/stanford-ribonanza-rna-folding/discussion/442089 hope it will helpful … Sir @danishmirza0321
", "votes": 0, "replies": [] }, { "id": 2453212, "author_name": "", "author_url": "", "post_date": "2023-09-23T21:49:11.247000", "content": "", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "2434764": "Reactivity is the data result of chemical probing experiments where the folded molecules (sequences) are exposed to a reagent (2A3 or DMS) that attacks the molecule. The bases that are paired are protected from exposure and have no reactivity. The bases that are unpaired are exposed and have high reactivity. Naturally, the process and data are not as exact as my description implies.\n\nSomething that hasn't been noted elsewhere yet is that 2A3 reacts with all bases whereas DMS only reacts with A and C. See [Wikipedia DMS](https://en.wikipedia.org/wiki/Dimethyl_sulfate) under \"Reactions with nucleic acids\".\n\nChemical probing is widely used in RNA structure research due to its lower cost and shorter timeframe compared with the 3D structure determination methods used to generate PDB files. Cryo-EM is particularly challenging with RNA due to its small size.", "2435006": "> Isn't the RNA sequence enough to determine the structure using dynamic programming?\n\nUnfortunately, no! Consider the very short sequence GGAAAACC. This could have the following sets of pairs:\n* 1-8, 2-7\n* 1-8\n* 2-7\n* 1-7\n* 2-8\n* no pairs at all\n\nConventionally structure is predicted using dynamic programming and comes up with the best option based on measured thermodynamic properties, but those predicted structures still have issues - and conventionally don't take into account pseudoknots.\n\n> Do we have to predict the reactivity which are null valued in the training set?\n\nAccording to the sample submission and information on the home page, it looks like it. I'd _assume_ that when data is unavailable it wouldn't be factored into the scoring, but I'll leave that to be confirmed by someone who has more details on the scoring", "2434448": "Can someone brief me the biological part of the problem. How does reactivity relates to the structure of the RNA? Isn't the RNA sequence enough to determine the structure using dynamic programming? Do we have to predict the reactivity which are null valued in the training set?", "2449555": "https://www.kaggle.com/competitions/stanford-ribonanza-rna-folding/discussion/442089 hope it will helpful ... Sir @danishmirza0321 ", "2453212": "" } }