For a molecule, are all possible pairs included in train or test? For instance, for type 2JHH, are all pairs of hydrogen atoms separated by 2 bonds included in train.csv or test.csv ?
", "messages": [ { "id": "571939", "postDate": "07/10/2019 08:55:58", "content": "For a molecule, are all possible pairs included in train or test? For instance, for type 2JHH, are all pairs of hydrogen atoms separated by 2 bonds included in train.csv or test.csv ?
", "rawMarkdown": "For a molecule, are all possible pairs included in train or test? For instance, for type 2JHH, are all pairs of hydrogen atoms separated by 2 bonds included in train.csv or test.csv ?", "votes": null }, { "id": "571949", "postDate": "07/10/2019 09:03:35", "content": "in Data Description:\n*For this competition, you will not be predicting all the atom pairs in each molecule rather, you will only need to predict the pairs that are explicitly listed in the train and test files. For example, some molecules contain Fluorine (F), but you will not be predicting the scalar coupling constant for any pair that includes F.\n*
", "rawMarkdown": "in Data Description:\n*For this competition, you will not be predicting all the atom pairs in each molecule rather, you will only need to predict the pairs that are explicitly listed in the train and test files. For example, some molecules contain Fluorine (F), but you will not be predicting the scalar coupling constant for any pair that includes F.\n*", "votes": null }, { "id": "571952", "postDate": "07/10/2019 09:08:32", "content": "I read that. I get that no pair with F are included. But it does not say if all other pairs are included.
", "rawMarkdown": "I read that. I get that no pair with F are included. But it does not say if all other pairs are included.", "votes": null }, { "id": "572138", "postDate": "07/10/2019 14:23:25", "content": "I verified this to be true for 1JHC, 1JHC, 2JHH, 2JHC and 2JHN. \nThe bonds I calculated using openbabel and the molecules structures xyz files. The numbers equal the corresponding sum from train and test files. (Every molecule in the structure files are either in train or test.)\nUpdate: \nThis is also the case for 3JHH, but not 3JHC or 3JHN. The reasons are explained by Kyle Peter in his answer. \nI created a pandas dataframe df_bonds with (molecule_name, atom_index_0, atom_index_1) (and some additional information) with bonds determined by openbabel. Through joins one can construct the candidates for the 2JHH, 2JHC,.. couplings. For 1JHC, 1JHN, 2JHH, 2JHN and 3JHH no additional reduction is required - for this sample train and test data - but this is required for 3JHC and 3JHN.
\n\n2nd update:\nI completed the check for 3JHC and 3JHN. An atom tuple is of type nJHx if and only if the shortest path (through bonds) between the starting 'H' molecule and the ending x-molecule (x = 'H', 'C' or 'N') is of length n. For nJHH there is the additional condition of atom_index_0 < atom_index_1
", "rawMarkdown": "I verified this to be true for 1JHC, 1JHC, 2JHH, 2JHC and 2JHN. \nThe bonds I calculated using openbabel and the molecules structures xyz files. The numbers equal the corresponding sum from train and test files. (Every molecule in the structure files are either in train or test.)\nUpdate: \nThis is also the case for 3JHH, but not 3JHC or 3JHN. The reasons are explained by Kyle Peter in his answer. \nI created a pandas dataframe df\\_bonds with (molecule\\_name, atom\\_index_0, atom\\_index_1) (and some additional information) with bonds determined by openbabel. Through joins one can construct the candidates for the 2JHH, 2JHC,.. couplings. For 1JHC, 1JHN, 2JHH, 2JHN and 3JHH no additional reduction is required - for this sample train and test data - but this is required for 3JHC and 3JHN. \n\n2nd update:\nI completed the check for 3JHC and 3JHN. An atom tuple is of type nJHx if and only if the shortest path (through bonds) between the starting 'H' molecule and the ending x-molecule (x = 'H', 'C' or 'N') is of length n. For nJHH there is the additional condition of atom\\_index\\_0 < atom\\_index\\_1", "votes": null }, { "id": "572141", "postDate": "07/10/2019 14:26:28", "content": "Thanks a lot! I hope someone also checked the 3J types.
", "rawMarkdown": "Thanks a lot! I hope someone also checked the 3J types.", "votes": null }, { "id": "572436", "postDate": "07/11/2019 00:15:30", "content": "For coupling to occur, the atoms need to be chemically in equivalent. So, for 2JHH and 3JHH coupling, there must be at least one example of the two hydrogens being chemically equivalent (same chemical shift) and thus that pair wouldn't be in the training set which would make me believe that the answer to your question is no.
", "rawMarkdown": "For coupling to occur, the atoms need to be chemically in equivalent. So, for 2JHH and 3JHH coupling, there must be at least one example of the two hydrogens being chemically equivalent (same chemical shift) and thus that pair wouldn't be in the training set which would make me believe that the answer to your question is no.", "votes": null }, { "id": "572538", "postDate": "07/11/2019 03:53:42", "content": "\n\n\nFor coupling to occur, the atoms need to be chemically in equivalent.
\n
I don't understand this at all.
\n\nFor 1JHC, the two atoms, H and C are not chemically equivalent. Yet there is a coupling obviously.
", "rawMarkdown": "> For coupling to occur, the atoms need to be chemically in equivalent. \n\nI don't understand this at all.\n\nFor 1JHC, the two atoms, H and C are not chemically equivalent. Yet there is a coupling obviously.", "votes": null }, { "id": "572773", "postDate": "07/11/2019 11:44:32", "content": "\"Coupling is only observed between protons that are magnetically non-equivalent\"\nhttp://chemistry.syr.edu/totah/che575/support/3a1/7-1.JNMR.pdf
", "rawMarkdown": "\"Coupling is only observed between protons that are magnetically non-equivalent\"\nhttp://chemistry.syr.edu/totah/che575/support/3a1/7-1.JNMR.pdf", "votes": null }, { "id": "572776", "postDate": "07/11/2019 11:48:28", "content": "Also, \"Chemically equivalent atoms (those with the same magnetic environment) don't couple together!\" from https://www.kaggle.com/archaizzard/simple-basics-of-nmr
", "rawMarkdown": "Also, \"Chemically equivalent atoms (those with the same magnetic environment) don't couple together!\" from https://www.kaggle.com/archaizzard/simple-basics-of-nmr", "votes": null }, { "id": "572815", "postDate": "07/11/2019 12:35:19", "content": "I read equivalent in your sentence, which is why i asked. I didn't thought that in equivalent was one word ;)
", "rawMarkdown": "I read *equivalent* in your sentence, which is why i asked. I didn't thought that *in equivalent* was one word ;)", "votes": null }, { "id": "572859", "postDate": "07/11/2019 13:41:18", "content": "@petersk20 Something puzzles me still. The first molecule is methane with 4 hydrogen atoms that are equivalent. All pairs 2JHH are in train data. And their coupling is not near 0.
", "rawMarkdown": "petersk20 Something puzzles me still. The first molecule is methane with 4 hydrogen atoms that are equivalent. All pairs 2JHH are in train data. And their coupling is not near 0.", "votes": null }, { "id": "572896", "postDate": "07/11/2019 14:44:27", "content": "After further research, I see that coupling between magnetically equivalent atoms can not be observed on an NMR spectrum. However, I think they still couple. The organizers must have used a method of calculating the coupling that could detect the constant even between magnetically equivalent atoms.
", "rawMarkdown": "After further research, I see that coupling between magnetically equivalent atoms can not be observed on an NMR spectrum. However, I think they still couple. The organizers must have used a method of calculating the coupling that could detect the constant even between magnetically equivalent atoms.", "votes": null }, { "id": "573008", "postDate": "07/11/2019 16:46:42", "content": "H / D isotope substitution method? (Sorry if I'm wrong - toooo far away from chemistry)))\nJНH -> ~6,5144JНD
", "rawMarkdown": "H / D isotope substitution method? (Sorry if I'm wrong - toooo far away from chemistry)))\nJНH -> ~6,5144JНD", "votes": null }, { "id": "577904", "postDate": "07/17/2019 07:23:09", "content": "I completed the description above also for 3J-couplings
", "rawMarkdown": "I completed the description above also for 3J-couplings", "votes": null }, { "id": "577918", "postDate": "07/17/2019 07:36:49", "content": "Thank you for the follow up. I resorted to computing bonds with openbabel since I asked.
", "rawMarkdown": "Thank you for the follow up. I resorted to computing bonds with openbabel since I asked.", "votes": null } ], "comments": [ { "id": 571949, "author_name": "fangwenjun", "author_url": "", "post_date": "07/10/2019 09:03:35", "content": "in Data Description:\n*For this competition, you will not be predicting all the atom pairs in each molecule rather, you will only need to predict the pairs that are explicitly listed in the train and test files. For example, some molecules contain Fluorine (F), but you will not be predicting the scalar coupling constant for any pair that includes F.\n*
", "votes": null, "replies": [ { "id": 571952, "author_name": "cpmpml", "author_url": "", "post_date": "07/10/2019 09:08:32", "content": "I read that. I get that no pair with F are included. But it does not say if all other pairs are included.
", "votes": null, "replies": [] } ] }, { "id": 572138, "author_name": "helgith", "author_url": "", "post_date": "07/10/2019 14:23:25", "content": "I verified this to be true for 1JHC, 1JHC, 2JHH, 2JHC and 2JHN. \nThe bonds I calculated using openbabel and the molecules structures xyz files. The numbers equal the corresponding sum from train and test files. (Every molecule in the structure files are either in train or test.)\nUpdate: \nThis is also the case for 3JHH, but not 3JHC or 3JHN. The reasons are explained by Kyle Peter in his answer. \nI created a pandas dataframe df_bonds with (molecule_name, atom_index_0, atom_index_1) (and some additional information) with bonds determined by openbabel. Through joins one can construct the candidates for the 2JHH, 2JHC,.. couplings. For 1JHC, 1JHN, 2JHH, 2JHN and 3JHH no additional reduction is required - for this sample train and test data - but this is required for 3JHC and 3JHN.
\n\n2nd update:\nI completed the check for 3JHC and 3JHN. An atom tuple is of type nJHx if and only if the shortest path (through bonds) between the starting 'H' molecule and the ending x-molecule (x = 'H', 'C' or 'N') is of length n. For nJHH there is the additional condition of atom_index_0 < atom_index_1
", "votes": null, "replies": [ { "id": 572141, "author_name": "cpmpml", "author_url": "", "post_date": "07/10/2019 14:26:28", "content": "Thanks a lot! I hope someone also checked the 3J types.
", "votes": null, "replies": [] }, { "id": 577904, "author_name": "helgith", "author_url": "", "post_date": "07/17/2019 07:23:09", "content": "I completed the description above also for 3J-couplings
", "votes": null, "replies": [] }, { "id": 577918, "author_name": "cpmpml", "author_url": "", "post_date": "07/17/2019 07:36:49", "content": "Thank you for the follow up. I resorted to computing bonds with openbabel since I asked.
", "votes": null, "replies": [] } ] }, { "id": 572436, "author_name": "petersk20", "author_url": "", "post_date": "07/11/2019 00:15:30", "content": "For coupling to occur, the atoms need to be chemically in equivalent. So, for 2JHH and 3JHH coupling, there must be at least one example of the two hydrogens being chemically equivalent (same chemical shift) and thus that pair wouldn't be in the training set which would make me believe that the answer to your question is no.
", "votes": null, "replies": [ { "id": 572538, "author_name": "cpmpml", "author_url": "", "post_date": "07/11/2019 03:53:42", "content": "\n\n\nFor coupling to occur, the atoms need to be chemically in equivalent.
\n
I don't understand this at all.
\n\nFor 1JHC, the two atoms, H and C are not chemically equivalent. Yet there is a coupling obviously.
", "votes": null, "replies": [] }, { "id": 572773, "author_name": "petersk20", "author_url": "", "post_date": "07/11/2019 11:44:32", "content": "\"Coupling is only observed between protons that are magnetically non-equivalent\"\nhttp://chemistry.syr.edu/totah/che575/support/3a1/7-1.JNMR.pdf
", "votes": null, "replies": [] }, { "id": 572776, "author_name": "petersk20", "author_url": "", "post_date": "07/11/2019 11:48:28", "content": "Also, \"Chemically equivalent atoms (those with the same magnetic environment) don't couple together!\" from https://www.kaggle.com/archaizzard/simple-basics-of-nmr
", "votes": null, "replies": [] }, { "id": 572815, "author_name": "cpmpml", "author_url": "", "post_date": "07/11/2019 12:35:19", "content": "I read equivalent in your sentence, which is why i asked. I didn't thought that in equivalent was one word ;)
", "votes": null, "replies": [] }, { "id": 572859, "author_name": "cpmpml", "author_url": "", "post_date": "07/11/2019 13:41:18", "content": "@petersk20 Something puzzles me still. The first molecule is methane with 4 hydrogen atoms that are equivalent. All pairs 2JHH are in train data. And their coupling is not near 0.
", "votes": null, "replies": [] }, { "id": 572896, "author_name": "petersk20", "author_url": "", "post_date": "07/11/2019 14:44:27", "content": "After further research, I see that coupling between magnetically equivalent atoms can not be observed on an NMR spectrum. However, I think they still couple. The organizers must have used a method of calculating the coupling that could detect the constant even between magnetically equivalent atoms.
", "votes": null, "replies": [] }, { "id": 573008, "author_name": "kyakovlev", "author_url": "", "post_date": "07/11/2019 16:46:42", "content": "H / D isotope substitution method? (Sorry if I'm wrong - toooo far away from chemistry)))\nJНH -> ~6,5144JНD
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "571939": "For a molecule, are all possible pairs included in train or test? For instance, for type 2JHH, are all pairs of hydrogen atoms separated by 2 bonds included in train.csv or test.csv ?", "571949": "in Data Description:\n*For this competition, you will not be predicting all the atom pairs in each molecule rather, you will only need to predict the pairs that are explicitly listed in the train and test files. For example, some molecules contain Fluorine (F), but you will not be predicting the scalar coupling constant for any pair that includes F.\n*", "571952": "I read that. I get that no pair with F are included. But it does not say if all other pairs are included.", "572138": "I verified this to be true for 1JHC, 1JHC, 2JHH, 2JHC and 2JHN. \nThe bonds I calculated using openbabel and the molecules structures xyz files. The numbers equal the corresponding sum from train and test files. (Every molecule in the structure files are either in train or test.)\nUpdate: \nThis is also the case for 3JHH, but not 3JHC or 3JHN. The reasons are explained by Kyle Peter in his answer. \nI created a pandas dataframe df\\_bonds with (molecule\\_name, atom\\_index_0, atom\\_index_1) (and some additional information) with bonds determined by openbabel. Through joins one can construct the candidates for the 2JHH, 2JHC,.. couplings. For 1JHC, 1JHN, 2JHH, 2JHN and 3JHH no additional reduction is required - for this sample train and test data - but this is required for 3JHC and 3JHN. \n\n2nd update:\nI completed the check for 3JHC and 3JHN. An atom tuple is of type nJHx if and only if the shortest path (through bonds) between the starting 'H' molecule and the ending x-molecule (x = 'H', 'C' or 'N') is of length n. For nJHH there is the additional condition of atom\\_index\\_0 < atom\\_index\\_1", "572141": "Thanks a lot! I hope someone also checked the 3J types.", "572436": "For coupling to occur, the atoms need to be chemically in equivalent. So, for 2JHH and 3JHH coupling, there must be at least one example of the two hydrogens being chemically equivalent (same chemical shift) and thus that pair wouldn't be in the training set which would make me believe that the answer to your question is no.", "572538": "> For coupling to occur, the atoms need to be chemically in equivalent. \n\nI don't understand this at all.\n\nFor 1JHC, the two atoms, H and C are not chemically equivalent. Yet there is a coupling obviously.", "572773": "\"Coupling is only observed between protons that are magnetically non-equivalent\"\nhttp://chemistry.syr.edu/totah/che575/support/3a1/7-1.JNMR.pdf", "572776": "Also, \"Chemically equivalent atoms (those with the same magnetic environment) don't couple together!\" from https://www.kaggle.com/archaizzard/simple-basics-of-nmr", "572815": "I read *equivalent* in your sentence, which is why i asked. I didn't thought that *in equivalent* was one word ;)", "572859": "petersk20 Something puzzles me still. The first molecule is methane with 4 hydrogen atoms that are equivalent. All pairs 2JHH are in train data. And their coupling is not near 0.", "572896": "After further research, I see that coupling between magnetically equivalent atoms can not be observed on an NMR spectrum. However, I think they still couple. The organizers must have used a method of calculating the coupling that could detect the constant even between magnetically equivalent atoms.", "573008": "H / D isotope substitution method? (Sorry if I'm wrong - toooo far away from chemistry)))\nJНH -> ~6,5144JНD", "577904": "I completed the description above also for 3J-couplings", "577918": "Thank you for the follow up. I resorted to computing bonds with openbabel since I asked." }, "source": "meta" }