Please use this thread to list the external data you are planning to use. Once a data source has been posted the first time, it does not need to be posted again by other competitors planning to use the data. You must post here one week prior to competition close.
\n\nPlease refer to the Rules page on important restrictions on what external data cannot be used for this competition.
", "messages": [ { "id": 539280, "postDate": "2019-05-29T20:14:42.870Z", "content": "Please use this thread to list the external data you are planning to use. Once a data source has been posted the first time, it does not need to be posted again by other competitors planning to use the data. You must post here one week prior to competition close.
\n\nPlease refer to the Rules page on important restrictions on what external data cannot be used for this competition.
", "rawMarkdown": "Please use this thread to list the external data you are planning to use. Once a data source has been posted the first time, it does not need to be posted again by other competitors planning to use the data. You must post here one week prior to competition close.\n\nPlease refer to the Rules page on important restrictions on what external data *cannot* be used for this competition.", "votes": 34 }, { "id": 539314, "postDate": "2019-05-29T22:10:53.307Z", "content": "Sanity Check:\n1. http://nmrwiki.org/wiki/index.php?title=Scalar_coupling_values\n2. http://sopnmr.ucsd.edu/coupling.htm
\n\nUseful Starter links for background Knowledge:\n1. https://www.khanacademy.org/science/organic-chemistry/spectroscopy-jay/proton-nmr/v/coupling-constant
", "rawMarkdown": "Sanity Check:\n1. http://nmrwiki.org/wiki/index.php?title=Scalar_coupling_values\n2. http://sopnmr.ucsd.edu/coupling.htm\n\nUseful Starter links for background Knowledge:\n1. https://www.khanacademy.org/science/organic-chemistry/spectroscopy-jay/proton-nmr/v/coupling-constant", "votes": 26, "replies": [ { "id": 547860, "postDate": "2019-06-08T11:47:39.163Z", "content": "Thank you for the useful Khan link. I recommend starting even at the first lesson.
", "rawMarkdown": "Thank you for the useful Khan link. I recommend starting even at the first lesson." } ] }, { "id": 539371, "postDate": "2019-05-30T01:59:42.167Z", "content": "This may be useful: https://github.com/lmmentel/awesome-python-chemistry
", "rawMarkdown": "This may be useful: https://github.com/lmmentel/awesome-python-chemistry", "votes": 11 }, { "id": 540386, "postDate": "2019-05-31T12:49:23.803Z", "content": "Hi inversion, I am not a chemist or physicist so I have no idea about these QM datasets, but I want to ask - what kind of information is unique to the QM datasets?
\n\nFor example, we are given dipole_moments, magnetic_shielding_tensors, mulliken_charges, potential_energy and scalar_coupling_contributions for our training samples. I am assuming that this data is unique to the QM datasets, which is why it was not provided to us for our test data? I would like to confirm this is the case so that I don't worry about there being important data out there that would hold back my predictions. Thanks.
", "rawMarkdown": "Hi inversion, I am not a chemist or physicist so I have no idea about these QM datasets, but I want to ask - what kind of information is unique to the QM datasets?\n\nFor example, we are given dipole_moments, magnetic_shielding_tensors, mulliken_charges, potential_energy and scalar_coupling_contributions for our training samples. I am assuming that this data is unique to the QM datasets, which is why it was not provided to us for our test data? I would like to confirm this is the case so that I don't worry about there being important data out there that would hold back my predictions. Thanks.", "votes": 12, "replies": [ { "id": 544172, "postDate": "2019-06-05T08:28:58.653Z", "content": "I have the same question
", "rawMarkdown": "I have the same question" } ] }, { "id": 558475, "postDate": "2019-06-22T11:31:21.763Z", "content": "Can we use rulers ;)
", "rawMarkdown": "Can we use rulers ;)", "votes": 8, "replies": [ { "id": 569476, "postDate": "2019-07-06T18:21:04.533Z", "content": "I bring the ruler, you bring the paper! How about that?
", "rawMarkdown": "I bring the ruler, you bring the paper! How about that?", "votes": 3 }, { "id": 569543, "postDate": "2019-07-06T21:48:05.703Z", "content": "1st place team started here ^
", "rawMarkdown": "1st place team started here ^", "votes": 3 } ] }, { "id": 548692, "postDate": "2019-06-09T18:00:22.810Z", "content": "Can we use SMILES and RDKit?
", "rawMarkdown": "Can we use SMILES and RDKit?", "votes": 5 }, { "id": 604167, "postDate": "2019-08-21T05:28:57.473Z", "content": "https://pubchem.ncbi.nlm.nih.gov/
", "rawMarkdown": "https://pubchem.ncbi.nlm.nih.gov/\n", "votes": 1 }, { "id": 603809, "postDate": "2019-08-20T17:21:21.490Z", "content": "Hello, I am using atomic size and electronegativity values and atomic covalent radius values from these sources.\nThe electronegativity values source may have been posted already, and supposedly no two lists of atomic radii have the same values, so I am going with this one.
\n\nhttps://www.lenntech.com/periodic-chart-elements/electronegativity.htm\nhttps://en.wikipedia.org/wiki/Covalent_radius
", "rawMarkdown": "Hello, I am using atomic size and electronegativity values and atomic covalent radius values from these sources.\nThe electronegativity values source may have been posted already, and supposedly no two lists of atomic radii have the same values, so I am going with this one.\n\nhttps://www.lenntech.com/periodic-chart-elements/electronegativity.htm\nhttps://en.wikipedia.org/wiki/Covalent_radius\n", "votes": 1 }, { "id": 590304, "postDate": "2019-08-02T03:26:09.303Z", "content": "I saw a data visualisation kernel use electronegativity and atomic size values. If I find some benefit using them, is it alright to use these values for prediction?
", "rawMarkdown": "I saw a data visualisation kernel use electronegativity and atomic size values. If I find some benefit using them, is it alright to use these values for prediction?", "votes": 1 }, { "id": 546563, "postDate": "2019-06-06T17:27:46.553Z", "content": "What about methods like GAFF or CGenFF that are widely used in the molecular modeling roam to generate atomic partial charges, bonds and other force-field parameters?
", "rawMarkdown": "What about methods like GAFF or CGenFF that are widely used in the molecular modeling roam to generate atomic partial charges, bonds and other force-field parameters?", "votes": 1 }, { "id": 546416, "postDate": "2019-06-06T15:14:32.957Z", "content": "Can we use SMILES (which is not from quantum mechanics computations) in GDB9-14(QM9)?
", "rawMarkdown": "Can we use SMILES (which is not from quantum mechanics computations) in GDB9-14(QM9)?", "votes": 1 }, { "id": 539720, "postDate": "2019-05-30T12:38:50.990Z", "content": "Can someone clarify this excerpt from the rules please? (Emphasis mine:)
\n\n\n\n\nYou may use data other than the Competition Data (“External Data”) to develop and test your models and \n Submissions excluding data from quantum mechanics computations made specifically for the competition, > and excluding data from quantum mechanics computations on the molecules from the training and test set, > including the QM7 and QM9 data sets.
\n
Does this imply that the QM9 datasets are included in the preceding two exclusions, and are thus ineligible for use as external data sources? Or is any external data OK as long as it is not the output of a QM calculation on any molecules in the test and/or training sets?
", "rawMarkdown": "Can someone clarify this excerpt from the rules please? (Emphasis mine:)\n\n> You may use data other than the Competition Data (“External Data”) to develop and test your models and \n> Submissions **excluding** data from quantum mechanics computations made specifically for the competition, > and **excluding** data from quantum mechanics computations on the molecules from the training and test set, > **including** the QM7 and QM9 data sets.\n\nDoes this imply that the QM9 datasets are included in the preceding two exclusions, and are thus ineligible for use as external data sources? Or is any external data OK as long as it is **not** the output of a QM calculation on any molecules in the test and/or training sets?", "votes": 1, "replies": [ { "id": 539891, "postDate": "2019-05-30T16:58:26.707Z", "content": "You can't use the QM datasets.
", "rawMarkdown": "You can't use the QM datasets.", "votes": 7 }, { "id": 549008, "postDate": "2019-06-10T07:36:00.110Z", "content": "@inversion would you be kind enough to review this and perhaps comment? thanks!\nhttps://www.kaggle.com/c/champs-scalar-coupling/discussion/95047#latest-548611
", "rawMarkdown": "@inversion would you be kind enough to review this and perhaps comment? thanks!\nhttps://www.kaggle.com/c/champs-scalar-coupling/discussion/95047#latest-548611", "votes": 1 } ] }, { "id": 583093, "postDate": "2019-07-24T02:48:37.437Z", "content": "There are a number of items posted that might seem borderline for use per the rules - are we going to get any feedback either approving or nixing some of the available libraries - many of which are highly voted right now.
", "rawMarkdown": "There are a number of items posted that might seem borderline for use per the rules - are we going to get any feedback either approving or nixing some of the available libraries - many of which are highly voted right now.", "votes": 2 }, { "id": 552902, "postDate": "2019-06-14T19:23:07.823Z", "content": "Learning Invariant Representations of Molecules for Atomization Energy Prediction
", "rawMarkdown": "[Learning Invariant Representations of Molecules for Atomization Energy Prediction](https://papers.nips.cc/paper/4830-learning-invariant-representations-of-molecules-for-atomization-energy-prediction)", "votes": 2 }, { "id": 539513, "postDate": "2019-05-30T06:26:30.533Z", "rawMarkdown": "", "votes": 2 }, { "id": 543921, "postDate": "2019-06-04T23:59:17.877Z", "content": "This might be useful http://www.gromacs.org/
", "rawMarkdown": "This might be useful http://www.gromacs.org/", "replies": [ { "id": 546583, "postDate": "2019-06-06T17:42:16.963Z", "content": "GROMACS is a program to run Molecular Dynamics Simulations of biological molecules such as proteins. The force-fields used there are usually for poly-peptides (proteins). In this competition we have small organic molecules. You would have to generate force-field parameters for each molecule. A parameterization usually involves quantum-mechanical optimization for structure optimization and partial charge calculation. There are frameworks that ommit these kind of computations though. Such as GAFF. GAFF is made to generate QM partial charges without necessarily performing QM calculations. In order to use GAFF you need to install Antechamber . Which is not part of ambertools. You can download it or install it via conda as well:
\n\n conda install ambertools -c ambermd\nAlso, OpenBabel the chemical file format converter comes to my mind. I think it generates Gasteiger charges by default, which are considered rather inaccurate for advanced simulations (same for Mulliken charges). It is contained in Linux repositories
\n\nsudo apt install openbabel\nThere is also RDkit that can be used to calculate Gasteiger Charges and many other molecular descriptors. You can install it with
\n\nconda install -c rdkit rdkit\nGiba Features : https://www.kaggle.com/scaomath/giba-molecular-features\nQM9 data : https://www.kaggle.com/zaharch/quantum-machine-9-qm9
", "rawMarkdown": "Giba Features : https://www.kaggle.com/scaomath/giba-molecular-features\nQM9 data : https://www.kaggle.com/zaharch/quantum-machine-9-qm9" }, { "id": 604841, "postDate": "2019-08-21T20:21:37.580Z", "content": "I may or may not use some of the below public datasets, resources and/or software packages by the end of the competition:
\n\nPublic resources and datasets:
\n\nSoftware packages:
\n\nI might use the QM7 and QM9 datasets, as this was allowed in the rule change.
", "rawMarkdown": "I might use the QM7 and QM9 datasets, as this was allowed in the rule change." }, { "id": 604315, "postDate": "2019-08-21T09:03:13.363Z", "content": "bonds between atoms\nhttps://www.kaggle.com/asauve/dataset-with-number-of-bonds-between-atoms/output
", "rawMarkdown": "bonds between atoms\nhttps://www.kaggle.com/asauve/dataset-with-number-of-bonds-between-atoms/output" }, { "id": 604049, "postDate": "2019-08-21T01:24:02.897Z", "content": "Bond energy, bond length data\n- https://chem.libretexts.org/Bookshelves/Physical_and_Theoretical_Chemistry_Textbook_Maps/Supplemental_Modules_(Physical_and_Theoretical_Chemistry)/Chemical_Bonding/Fundamentals_of_Chemical_Bonding/Chemical_Bonds/Bond_Lengths_and_Energies
\n\nQM9 in deepchem python package:
\n\nhttps://github.com/deepchem/deepchem/blob/master/deepchem/molnet/load_function/qm9_datasets.py
\n\n(about deepchem visit: https://github.com/deepchem/deepchem)
", "rawMarkdown": "QM9 in deepchem python package:\n\nhttps://github.com/deepchem/deepchem/blob/master/deepchem/molnet/load_function/qm9_datasets.py\n\n(about deepchem visit: https://github.com/deepchem/deepchem)" }, { "id": 603128, "postDate": "2019-08-19T22:17:34.030Z", "content": "https://singroup.github.io/dscribe/
", "rawMarkdown": "https://singroup.github.io/dscribe/" }, { "id": 602164, "postDate": "2019-08-18T17:44:00.693Z", "content": "https://en.wikipedia.org/wiki/Nitrogen-15_nuclear_magnetic_resonance_spectroscopy
\n\nphysical properties
", "rawMarkdown": "https://en.wikipedia.org/wiki/Nitrogen-15_nuclear_magnetic_resonance_spectroscopy\n\nphysical properties" }, { "id": 599647, "postDate": "2019-08-15T08:03:28.997Z", "content": "Used atomic properties like electronegativity, weight, size etc of the given atoms (H, C, N, O and F)
", "rawMarkdown": "Used atomic properties like electronegativity, weight, size etc of the given atoms (H, C, N, O and F)" }, { "id": 597664, "postDate": "2019-08-12T16:22:24.730Z", "content": "Basic information about the atoms like number of electrons, electro-negativity values etc.
", "rawMarkdown": "Basic information about the atoms like number of electrons, electro-negativity values etc." }, { "id": 591487, "postDate": "2019-08-03T19:32:34.733Z", "content": "https://en.wikipedia.org/wiki/Atomic_radii_of_the_elements_(data_page)
", "rawMarkdown": "https://en.wikipedia.org/wiki/Atomic_radii_of_the_elements_(data_page)" }, { "id": 586833, "postDate": "2019-07-29T19:03:56.497Z", "content": "Publicly available Knodle (https://team.inria.fr/nano-d/software/knodle/) -- may one use it?\nNot a quantum chemistry package, just a small atom type classification binary.
", "rawMarkdown": "Publicly available Knodle (https://team.inria.fr/nano-d/software/knodle/) -- may one use it?\nNot a quantum chemistry package, just a small atom type classification binary." }, { "id": 585394, "postDate": "2019-07-27T11:16:45.770Z", "content": "", "rawMarkdown": "http://zinc15.docking.org" }, { "id": 584856, "postDate": "2019-07-26T15:26:19.290Z", "content": "Bond lengths and energies
\n\nhttp://www.wiredchemist.com/chemistry/data/bond_energies_lengths.html\nhttp://hydra.vcp.monash.edu.au/modules/mod2/bondlen.html\nhttps://chem.libretexts.org/Bookshelves/Physical_and_Theoretical_Chemistry_Textbook_Maps/Supplemental_Modules_(Physical_and_Theoretical_Chemistry)/Chemical_Bonding/Fundamentals_of_Chemical_Bonding/Chemical_Bonds/Bond_Lengths_and_Energies
", "rawMarkdown": "Bond lengths and energies\n\nhttp://www.wiredchemist.com/chemistry/data/bond_energies_lengths.html\nhttp://hydra.vcp.monash.edu.au/modules/mod2/bondlen.html\nhttps://chem.libretexts.org/Bookshelves/Physical_and_Theoretical_Chemistry_Textbook_Maps/Supplemental_Modules_(Physical_and_Theoretical_Chemistry)/Chemical_Bonding/Fundamentals_of_Chemical_Bonding/Chemical_Bonds/Bond_Lengths_and_Energies" }, { "id": 573876, "postDate": "2019-07-12T22:06:50.413Z", "content": "Here is the output of my kernel for angle and dihedral calculations.
", "rawMarkdown": "[Here](https://www.kaggle.com/soerendip/angle-and-dihedral-for-the-champs-structures) is the output of my [kernel](https://www.kaggle.com/soerendip/calculate-angles-and-dihedrals-with-networkx) for angle and dihedral calculations.\n" }, { "id": 559915, "postDate": "2019-06-24T19:09:43.747Z", "content": "supplemental data from here: https://pubs.acs.org/doi/suppl/10.1021/acs.jctc.7b00577/suppl_file/ct7b00577_si_001.pdf
", "rawMarkdown": "supplemental data from here: https://pubs.acs.org/doi/suppl/10.1021/acs.jctc.7b00577/suppl_file/ct7b00577_si_001.pdf" }, { "id": 556824, "postDate": "2019-06-20T18:03:38.747Z", "content": "Will use QM9.
\n\nDoes the usage of other open-source libraries have to be disclosed? I see that people are sharing links to packages on this thread, but is it compulsory to do so? To be more specific, I am referring to packages that do not explicitly contain or make use of quantum chemical data.
", "rawMarkdown": "Will use QM9.\n\nDoes the usage of other open-source libraries have to be disclosed? I see that people are sharing links to packages on this thread, but is it compulsory to do so? To be more specific, I am referring to packages that do not explicitly contain or make use of quantum chemical data.", "replies": [ { "id": 558443, "postDate": "2019-06-22T09:58:44.677Z", "content": "@inversion says you can't use QM datasets. See comment above
", "rawMarkdown": "@inversion says you can't use QM datasets. See comment above", "votes": -2 }, { "id": 559659, "postDate": "2019-06-24T11:43:02.640Z", "content": "The rules were changed here: https://www.kaggle.com/c/champs-scalar-coupling/discussion/95496#latest-558861
", "rawMarkdown": "The rules were changed here: https://www.kaggle.com/c/champs-scalar-coupling/discussion/95496#latest-558861", "votes": 3 } ] }, { "id": 556152, "postDate": "2019-06-19T22:36:28.617Z", "content": "Openbabel.
", "rawMarkdown": "Openbabel." }, { "id": 556071, "postDate": "2019-06-19T20:01:21.763Z", "content": "How about the total energies of the involved atoms? Would it be allowed to calculate the energy of H, C, O, N, ... using e.g., GPAW?
", "rawMarkdown": "How about the total energies of the involved atoms? Would it be allowed to calculate the energy of H, C, O, N, ... using e.g., GPAW?" }, { "id": 551741, "postDate": "2019-06-13T04:21:22.990Z", "content": "As of this day forward I might be using QM7 and QM9 datasets.
", "rawMarkdown": "As of this day forward I might be using QM7 and QM9 datasets." }, { "id": 549918, "postDate": "2019-06-11T06:26:22.290Z", "content": "https://github.com/samoturk/mol2vec
", "rawMarkdown": "https://github.com/samoturk/mol2vec" }, { "id": 548603, "postDate": "2019-06-09T16:04:37.810Z", "content": "Is this allowed
\n\nhttps://www.kaggle.com/c/champs-scalar-coupling/discussion/95047#latest-548595?
", "rawMarkdown": "Is this allowed \n\nhttps://www.kaggle.com/c/champs-scalar-coupling/discussion/95047#latest-548595?" }, { "id": 548323, "postDate": "2019-06-09T07:29:27.787Z", "content": "https://amp.readthedocs.io/en/latest/installation.html
\n\nhttps://singroup.github.io/dscribe/install.html
", "rawMarkdown": "https://amp.readthedocs.io/en/latest/installation.html\n\nhttps://singroup.github.io/dscribe/install.html" }, { "id": 547578, "postDate": "2019-06-07T22:16:43.207Z", "content": "https://mendeleev.readthedocs.io - for some basic features of elements
", "rawMarkdown": "https://mendeleev.readthedocs.io - for some basic features of elements" }, { "id": 543872, "postDate": "2019-06-04T22:40:43.577Z", "content": "Does measured NMR spectra are allowed as external data? Given it is publicly available, open data.
", "rawMarkdown": "Does measured NMR spectra are allowed as external data? Given it is publicly available, open data.\n" }, { "id": 604783, "postDate": "2019-08-21T19:21:55.843Z", "rawMarkdown": "", "isDeleted": true }, { "id": 604516, "postDate": "2019-08-21T13:23:43.247Z", "rawMarkdown": "", "isDeleted": true }, { "id": 602887, "postDate": "2019-08-19T16:00:40.567Z", "rawMarkdown": "", "isDeleted": true }, { "id": 580194, "postDate": "2019-07-19T20:17:09.270Z", "rawMarkdown": "", "isDeleted": true }, { "id": 573877, "postDate": "2019-07-12T22:07:22.440Z", "rawMarkdown": "", "isDeleted": true }, { "id": 569604, "postDate": "2019-07-07T02:23:56.357Z", "rawMarkdown": "", "isDeleted": true }, { "id": 559591, "postDate": "2019-06-24T10:04:54.507Z", "rawMarkdown": "", "isDeleted": true }, { "id": 552365, "postDate": "2019-06-13T20:00:41.243Z", "content": "Can we use all available methods in openbabel? Or are there some specific methods that are not allowed?
", "rawMarkdown": "Can we use all available methods in openbabel? Or are there some specific methods that are not allowed?", "votes": 8, "isDeleted": true }, { "id": 552117, "postDate": "2019-06-13T14:26:38Z", "rawMarkdown": "", "votes": 2, "isDeleted": true } ], "comments": [ { "id": 539314, "author_name": "Joel Lee", "author_url": "", "post_date": "2019-05-29T22:10:53.307000", "content": "Sanity Check:\n1. http://nmrwiki.org/wiki/index.php?title=Scalar_coupling_values\n2. http://sopnmr.ucsd.edu/coupling.htm
\n\nUseful Starter links for background Knowledge:\n1. https://www.khanacademy.org/science/organic-chemistry/spectroscopy-jay/proton-nmr/v/coupling-constant
", "votes": 26, "replies": [ { "id": 547860, "author_name": "Yohan Obadia", "author_url": "", "post_date": "2019-06-08T11:47:39.163000", "content": "Thank you for the useful Khan link. I recommend starting even at the first lesson.
", "votes": 0, "replies": [] } ] }, { "id": 539371, "author_name": "Rob Rose", "author_url": "", "post_date": "2019-05-30T01:59:42.167000", "content": "This may be useful: https://github.com/lmmentel/awesome-python-chemistry
", "votes": 11, "replies": [] }, { "id": 540386, "author_name": "Travis Black", "author_url": "", "post_date": "2019-05-31T12:49:23.803000", "content": "Hi inversion, I am not a chemist or physicist so I have no idea about these QM datasets, but I want to ask - what kind of information is unique to the QM datasets?
\n\nFor example, we are given dipole_moments, magnetic_shielding_tensors, mulliken_charges, potential_energy and scalar_coupling_contributions for our training samples. I am assuming that this data is unique to the QM datasets, which is why it was not provided to us for our test data? I would like to confirm this is the case so that I don't worry about there being important data out there that would hold back my predictions. Thanks.
", "votes": 12, "replies": [ { "id": 544172, "author_name": "Ranitha Mataraarachchi", "author_url": "", "post_date": "2019-06-05T08:28:58.653000", "content": "I have the same question
", "votes": 0, "replies": [] } ] }, { "id": 558475, "author_name": "Scirpus", "author_url": "", "post_date": "2019-06-22T11:31:21.763000", "content": "Can we use rulers ;)
", "votes": 8, "replies": [ { "id": 569476, "author_name": "CPMP", "author_url": "", "post_date": "2019-07-06T18:21:04.533000", "content": "I bring the ruler, you bring the paper! How about that?
", "votes": 3, "replies": [] }, { "id": 569543, "author_name": "Stanislav Blinov", "author_url": "", "post_date": "2019-07-06T21:48:05.703000", "content": "1st place team started here ^
", "votes": 3, "replies": [] } ] }, { "id": 548692, "author_name": "Melvin Yuen", "author_url": "", "post_date": "2019-06-09T18:00:22.810000", "content": "Can we use SMILES and RDKit?
", "votes": 5, "replies": [] }, { "id": 604167, "author_name": "PC Jimmmy", "author_url": "", "post_date": "2019-08-21T05:28:57.473000", "content": "https://pubchem.ncbi.nlm.nih.gov/
", "votes": 1, "replies": [] }, { "id": 603809, "author_name": "Aashish Ghosh", "author_url": "", "post_date": "2019-08-20T17:21:21.490000", "content": "Hello, I am using atomic size and electronegativity values and atomic covalent radius values from these sources.\nThe electronegativity values source may have been posted already, and supposedly no two lists of atomic radii have the same values, so I am going with this one.
\n\nhttps://www.lenntech.com/periodic-chart-elements/electronegativity.htm\nhttps://en.wikipedia.org/wiki/Covalent_radius
", "votes": 1, "replies": [] }, { "id": 590304, "author_name": "Aashish Ghosh", "author_url": "", "post_date": "2019-08-02T03:26:09.303000", "content": "I saw a data visualisation kernel use electronegativity and atomic size values. If I find some benefit using them, is it alright to use these values for prediction?
", "votes": 1, "replies": [] }, { "id": 546563, "author_name": "Klaus", "author_url": "", "post_date": "2019-06-06T17:27:46.553000", "content": "What about methods like GAFF or CGenFF that are widely used in the molecular modeling roam to generate atomic partial charges, bonds and other force-field parameters?
", "votes": 1, "replies": [] }, { "id": 546416, "author_name": "VarisaP", "author_url": "", "post_date": "2019-06-06T15:14:32.957000", "content": "Can we use SMILES (which is not from quantum mechanics computations) in GDB9-14(QM9)?
", "votes": 1, "replies": [] }, { "id": 539720, "author_name": "vonbraunbates", "author_url": "", "post_date": "2019-05-30T12:38:50.990000", "content": "Can someone clarify this excerpt from the rules please? (Emphasis mine:)
\n\n\n\n\nYou may use data other than the Competition Data (“External Data”) to develop and test your models and \n Submissions excluding data from quantum mechanics computations made specifically for the competition, > and excluding data from quantum mechanics computations on the molecules from the training and test set, > including the QM7 and QM9 data sets.
\n
Does this imply that the QM9 datasets are included in the preceding two exclusions, and are thus ineligible for use as external data sources? Or is any external data OK as long as it is not the output of a QM calculation on any molecules in the test and/or training sets?
", "votes": 1, "replies": [ { "id": 539891, "author_name": "inversion", "author_url": "", "post_date": "2019-05-30T16:58:26.707000", "content": "You can't use the QM datasets.
", "votes": 7, "replies": [] }, { "id": 549008, "author_name": "something4kag", "author_url": "", "post_date": "2019-06-10T07:36:00.110000", "content": "@inversion would you be kind enough to review this and perhaps comment? thanks!\nhttps://www.kaggle.com/c/champs-scalar-coupling/discussion/95047#latest-548611
", "votes": 1, "replies": [] } ] }, { "id": 583093, "author_name": "PC Jimmmy", "author_url": "", "post_date": "2019-07-24T02:48:37.437000", "content": "There are a number of items posted that might seem borderline for use per the rules - are we going to get any feedback either approving or nixing some of the available libraries - many of which are highly voted right now.
", "votes": 2, "replies": [] }, { "id": 552902, "author_name": "Shuhao Cao", "author_url": "", "post_date": "2019-06-14T19:23:07.823000", "content": "Learning Invariant Representations of Molecules for Atomization Energy Prediction
", "votes": 2, "replies": [] }, { "id": 539513, "author_name": "Jz", "author_url": "", "post_date": "2019-05-30T06:26:30.533000", "content": "", "votes": 2, "replies": [] }, { "id": 543921, "author_name": "Venky Krishnamani", "author_url": "", "post_date": "2019-06-04T23:59:17.877000", "content": "This might be useful http://www.gromacs.org/
", "votes": 0, "replies": [ { "id": 546583, "author_name": "Klaus", "author_url": "", "post_date": "2019-06-06T17:42:16.963000", "content": "GROMACS is a program to run Molecular Dynamics Simulations of biological molecules such as proteins. The force-fields used there are usually for poly-peptides (proteins). In this competition we have small organic molecules. You would have to generate force-field parameters for each molecule. A parameterization usually involves quantum-mechanical optimization for structure optimization and partial charge calculation. There are frameworks that ommit these kind of computations though. Such as GAFF. GAFF is made to generate QM partial charges without necessarily performing QM calculations. In order to use GAFF you need to install Antechamber . Which is not part of ambertools. You can download it or install it via conda as well:
\n\n conda install ambertools -c ambermd\nAlso, OpenBabel the chemical file format converter comes to my mind. I think it generates Gasteiger charges by default, which are considered rather inaccurate for advanced simulations (same for Mulliken charges). It is contained in Linux repositories
\n\nsudo apt install openbabel\nThere is also RDkit that can be used to calculate Gasteiger Charges and many other molecular descriptors. You can install it with
\n\nconda install -c rdkit rdkit\nGiba Features : https://www.kaggle.com/scaomath/giba-molecular-features\nQM9 data : https://www.kaggle.com/zaharch/quantum-machine-9-qm9
", "votes": 0, "replies": [] }, { "id": 604841, "author_name": "aagapi", "author_url": "", "post_date": "2019-08-21T20:21:37.580000", "content": "I may or may not use some of the below public datasets, resources and/or software packages by the end of the competition:
\n\nPublic resources and datasets:
\n\nSoftware packages:
\n\nI might use the QM7 and QM9 datasets, as this was allowed in the rule change.
", "votes": 0, "replies": [] }, { "id": 604315, "author_name": "Alexpartisan", "author_url": "", "post_date": "2019-08-21T09:03:13.363000", "content": "bonds between atoms\nhttps://www.kaggle.com/asauve/dataset-with-number-of-bonds-between-atoms/output
", "votes": 0, "replies": [] }, { "id": 604049, "author_name": "Ashesh.", "author_url": "", "post_date": "2019-08-21T01:24:02.897000", "content": "Bond energy, bond length data\n- https://chem.libretexts.org/Bookshelves/Physical_and_Theoretical_Chemistry_Textbook_Maps/Supplemental_Modules_(Physical_and_Theoretical_Chemistry)/Chemical_Bonding/Fundamentals_of_Chemical_Bonding/Chemical_Bonds/Bond_Lengths_and_Energies
\n\nQM9 in deepchem python package:
\n\nhttps://github.com/deepchem/deepchem/blob/master/deepchem/molnet/load_function/qm9_datasets.py
\n\n(about deepchem visit: https://github.com/deepchem/deepchem)
", "votes": 0, "replies": [] }, { "id": 603128, "author_name": "PC Jimmmy", "author_url": "", "post_date": "2019-08-19T22:17:34.030000", "content": "https://singroup.github.io/dscribe/
", "votes": 0, "replies": [] }, { "id": 602164, "author_name": "7rick03ligh7", "author_url": "", "post_date": "2019-08-18T17:44:00.693000", "content": "https://en.wikipedia.org/wiki/Nitrogen-15_nuclear_magnetic_resonance_spectroscopy
\n\nphysical properties
", "votes": 0, "replies": [] }, { "id": 599647, "author_name": "drzeus", "author_url": "", "post_date": "2019-08-15T08:03:28.997000", "content": "Used atomic properties like electronegativity, weight, size etc of the given atoms (H, C, N, O and F)
", "votes": 0, "replies": [] }, { "id": 597664, "author_name": "Krishnendu Sengupta", "author_url": "", "post_date": "2019-08-12T16:22:24.730000", "content": "Basic information about the atoms like number of electrons, electro-negativity values etc.
", "votes": 0, "replies": [] }, { "id": 591487, "author_name": "Thian Liang", "author_url": "", "post_date": "2019-08-03T19:32:34.733000", "content": "https://en.wikipedia.org/wiki/Atomic_radii_of_the_elements_(data_page)
", "votes": 0, "replies": [] }, { "id": 586833, "author_name": "Egor Marin", "author_url": "", "post_date": "2019-07-29T19:03:56.497000", "content": "Publicly available Knodle (https://team.inria.fr/nano-d/software/knodle/) -- may one use it?\nNot a quantum chemistry package, just a small atom type classification binary.
", "votes": 0, "replies": [] }, { "id": 585394, "author_name": "elmo", "author_url": "", "post_date": "2019-07-27T11:16:45.770000", "content": "", "votes": 0, "replies": [] }, { "id": 584856, "author_name": "Gautham Ganapathy", "author_url": "", "post_date": "2019-07-26T15:26:19.290000", "content": "Bond lengths and energies
\n\nhttp://www.wiredchemist.com/chemistry/data/bond_energies_lengths.html\nhttp://hydra.vcp.monash.edu.au/modules/mod2/bondlen.html\nhttps://chem.libretexts.org/Bookshelves/Physical_and_Theoretical_Chemistry_Textbook_Maps/Supplemental_Modules_(Physical_and_Theoretical_Chemistry)/Chemical_Bonding/Fundamentals_of_Chemical_Bonding/Chemical_Bonds/Bond_Lengths_and_Energies
", "votes": 0, "replies": [] }, { "id": 573876, "author_name": "", "author_url": "", "post_date": "2019-07-12T22:06:50.413000", "content": "", "votes": 0, "replies": [] }, { "id": 559915, "author_name": "", "author_url": "", "post_date": "2019-06-24T19:09:43.747000", "content": "", "votes": 0, "replies": [] }, { "id": 556824, "author_name": "", "author_url": "", "post_date": "2019-06-20T18:03:38.747000", "content": "", "votes": 0, "replies": [ { "id": 558443, "author_name": "", "author_url": "", "post_date": "2019-06-22T09:58:44.677000", "content": "", "votes": -2, "replies": [] }, { "id": 559659, "author_name": "", "author_url": "", "post_date": "2019-06-24T11:43:02.640000", "content": "", "votes": 3, "replies": [] } ] }, { "id": 556152, "author_name": "", "author_url": "", "post_date": "2019-06-19T22:36:28.617000", "content": "", "votes": 0, "replies": [] }, { "id": 556071, "author_name": "", "author_url": "", "post_date": "2019-06-19T20:01:21.763000", "content": "", "votes": 0, "replies": [] }, { "id": 551741, "author_name": "", "author_url": "", "post_date": "2019-06-13T04:21:22.990000", "content": "", "votes": 0, "replies": [] }, { "id": 549918, "author_name": "", "author_url": "", "post_date": "2019-06-11T06:26:22.290000", "content": "", "votes": 0, "replies": [] }, { "id": 548603, "author_name": "", "author_url": "", "post_date": "2019-06-09T16:04:37.810000", "content": "", "votes": 0, "replies": [] }, { "id": 548323, "author_name": "", "author_url": "", "post_date": "2019-06-09T07:29:27.787000", "content": "", "votes": 0, "replies": [] }, { "id": 547578, "author_name": "", "author_url": "", "post_date": "2019-06-07T22:16:43.207000", "content": "", "votes": 0, "replies": [] }, { "id": 543872, "author_name": "", "author_url": "", "post_date": "2019-06-04T22:40:43.577000", "content": "", "votes": 0, "replies": [] }, { "id": 604783, "author_name": "", "author_url": "", "post_date": "2019-08-21T19:21:55.843000", "content": "", "votes": 0, "replies": [] }, { "id": 604516, "author_name": "", "author_url": "", "post_date": "2019-08-21T13:23:43.247000", "content": "", "votes": 0, "replies": [] }, { "id": 602887, "author_name": "", "author_url": "", "post_date": "2019-08-19T16:00:40.567000", "content": "", "votes": 0, "replies": [] }, { "id": 580194, "author_name": "", "author_url": "", "post_date": "2019-07-19T20:17:09.270000", "content": "", "votes": 0, "replies": [] }, { "id": 573877, "author_name": "", "author_url": "", "post_date": "2019-07-12T22:07:22.440000", "content": "", "votes": 0, "replies": [] }, { "id": 569604, "author_name": "", "author_url": "", "post_date": "2019-07-07T02:23:56.357000", "content": "", "votes": 0, "replies": [] }, { "id": 559591, "author_name": "", "author_url": "", "post_date": "2019-06-24T10:04:54.507000", "content": "", "votes": 0, "replies": [] }, { "id": 552365, "author_name": "", "author_url": "", "post_date": "2019-06-13T20:00:41.243000", "content": "", "votes": 8, "replies": [] }, { "id": 552117, "author_name": "", "author_url": "", "post_date": "2019-06-13T14:26:38", "content": "", "votes": 2, "replies": [] } ], "raw_markdown_by_id": { "539280": "Please use this thread to list the external data you are planning to use. Once a data source has been posted the first time, it does not need to be posted again by other competitors planning to use the data. You must post here one week prior to competition close.\n\nPlease refer to the Rules page on important restrictions on what external data *cannot* be used for this competition.", "539314": "Sanity Check:\n1. http://nmrwiki.org/wiki/index.php?title=Scalar_coupling_values\n2. http://sopnmr.ucsd.edu/coupling.htm\n\nUseful Starter links for background Knowledge:\n1. https://www.khanacademy.org/science/organic-chemistry/spectroscopy-jay/proton-nmr/v/coupling-constant", "539371": "This may be useful: https://github.com/lmmentel/awesome-python-chemistry", "540386": "Hi inversion, I am not a chemist or physicist so I have no idea about these QM datasets, but I want to ask - what kind of information is unique to the QM datasets?\n\nFor example, we are given dipole_moments, magnetic_shielding_tensors, mulliken_charges, potential_energy and scalar_coupling_contributions for our training samples. I am assuming that this data is unique to the QM datasets, which is why it was not provided to us for our test data? I would like to confirm this is the case so that I don't worry about there being important data out there that would hold back my predictions. Thanks.", "558475": "Can we use rulers ;)", "548692": "Can we use SMILES and RDKit?", "604167": "https://pubchem.ncbi.nlm.nih.gov/\n", "603809": "Hello, I am using atomic size and electronegativity values and atomic covalent radius values from these sources.\nThe electronegativity values source may have been posted already, and supposedly no two lists of atomic radii have the same values, so I am going with this one.\n\nhttps://www.lenntech.com/periodic-chart-elements/electronegativity.htm\nhttps://en.wikipedia.org/wiki/Covalent_radius\n", "590304": "I saw a data visualisation kernel use electronegativity and atomic size values. If I find some benefit using them, is it alright to use these values for prediction?", "546563": "What about methods like GAFF or CGenFF that are widely used in the molecular modeling roam to generate atomic partial charges, bonds and other force-field parameters?", "546416": "Can we use SMILES (which is not from quantum mechanics computations) in GDB9-14(QM9)?", "539720": "Can someone clarify this excerpt from the rules please? (Emphasis mine:)\n\n> You may use data other than the Competition Data (“External Data”) to develop and test your models and \n> Submissions **excluding** data from quantum mechanics computations made specifically for the competition, > and **excluding** data from quantum mechanics computations on the molecules from the training and test set, > **including** the QM7 and QM9 data sets.\n\nDoes this imply that the QM9 datasets are included in the preceding two exclusions, and are thus ineligible for use as external data sources? Or is any external data OK as long as it is **not** the output of a QM calculation on any molecules in the test and/or training sets?", "583093": "There are a number of items posted that might seem borderline for use per the rules - are we going to get any feedback either approving or nixing some of the available libraries - many of which are highly voted right now.", "552902": "[Learning Invariant Representations of Molecules for Atomization Energy Prediction](https://papers.nips.cc/paper/4830-learning-invariant-representations-of-molecules-for-atomization-energy-prediction)", "539513": "", "543921": "This might be useful http://www.gromacs.org/", "607377": "Giba Features : https://www.kaggle.com/scaomath/giba-molecular-features\nQM9 data : https://www.kaggle.com/zaharch/quantum-machine-9-qm9", "604841": "I may or may not use some of the below public datasets, resources and/or software packages by the end of the competition:\n\nPublic resources and datasets:\n\n- https://gist.github.com/GoodmanSciences/c2dd862cd38f21b0ad36b8f96b4bf1ee\n- https://chem.libretexts.org/Ancillary_Materials/Reference/Reference_Tables/Atomic_and_Molecular_Properties/A3%3A_Covalent_Radii\n- https://en.wikipedia.org/wiki/Atomic_radii_of_the_elements_(data_page)\n- https://en.wikipedia.org/wiki/Covalent_radius\n- http://www.wiredchemist.com/chemistry/data/bond_energies_lengths.html\n- https://chem.libretexts.org/Bookshelves/Physical_and_Theoretical_Chemistry_Textbook_Maps/Supplemental_Modules_(Physical_and_Theoretical_Chemistry)/Chemical_Bonding/Fundamentals_of_Chemical_Bonding/Chemical_Bonds/Bond_Lengths_and_Energies\n- http://hydra.vcp.monash.edu.au/modules/mod2/bondlen.html\n- http://nmrwiki.org/wiki/index.php?title=Scalar_coupling_values\n- http://sopnmr.ucsd.edu/coupling.htm\n- https://pubchem.ncbi.nlm.nih.gov/\n- https://www.lenntech.com/periodic-chart-elements/electronegativity.htm\n- https://www.kaggle.com/soerendip/angle-and-dihedral-for-the-champs-structures\n- https://www.kaggle.com/soerendip/calculate-angles-and-dihedrals-with-networkx\n- https://www.kaggle.com/asauve/predicting-molecular-properties-bonds\n- https://www.kaggle.com/asauve/dataset-with-number-of-bonds-between-atoms/output\n- https://www.kaggle.com/chrisslz/makegraphinput\n- https://www.kaggle.com/bigironsphere/simple-molecular-geometry-features\n- QM7 and/or QM9 datasets\n\nSoftware packages:\n\n- rdkit: https://www.rdkit.org/\n- xyz2mol: https://github.com/jensengroup/xyz2mol\n- openbabel: https://openbabel.org\n- https://singroup.github.io/dscribe/\n- https://github.com/samoturk/mol2vec\n- https://mendeleev.readthedocs.io/en/latest/\n- https://amp.readthedocs.io/en/latest/introduction.html\n", "604659": "I might use the QM7 and QM9 datasets, as this was allowed in the rule change.", "604315": "bonds between atoms\nhttps://www.kaggle.com/asauve/dataset-with-number-of-bonds-between-atoms/output", "604049": "Bond energy, bond length data\n- https://chem.libretexts.org/Bookshelves/Physical_and_Theoretical_Chemistry_Textbook_Maps/Supplemental_Modules_(Physical_and_Theoretical_Chemistry)/Chemical_Bonding/Fundamentals_of_Chemical_Bonding/Chemical_Bonds/Bond_Lengths_and_Energies\n\n- http://www.wiredchemist.com/chemistry/data/bond_energies_lengths.html\n- https://en.wikipedia.org/wiki/Carbon%E2%80%93nitrogen_bond\n- Openbabel for some features for atoms and edges. http://openbabel.org/wiki/Main_Page", "603602": "QM9 in deepchem python package:\n\nhttps://github.com/deepchem/deepchem/blob/master/deepchem/molnet/load_function/qm9_datasets.py\n\n(about deepchem visit: https://github.com/deepchem/deepchem)", "603128": "https://singroup.github.io/dscribe/", "602164": "https://en.wikipedia.org/wiki/Nitrogen-15_nuclear_magnetic_resonance_spectroscopy\n\nphysical properties", "599647": "Used atomic properties like electronegativity, weight, size etc of the given atoms (H, C, N, O and F)", "597664": "Basic information about the atoms like number of electrons, electro-negativity values etc.", "591487": "https://en.wikipedia.org/wiki/Atomic_radii_of_the_elements_(data_page)", "586833": "Publicly available Knodle (https://team.inria.fr/nano-d/software/knodle/) -- may one use it?\nNot a quantum chemistry package, just a small atom type classification binary.", "585394": "http://zinc15.docking.org", "584856": "Bond lengths and energies\n\nhttp://www.wiredchemist.com/chemistry/data/bond_energies_lengths.html\nhttp://hydra.vcp.monash.edu.au/modules/mod2/bondlen.html\nhttps://chem.libretexts.org/Bookshelves/Physical_and_Theoretical_Chemistry_Textbook_Maps/Supplemental_Modules_(Physical_and_Theoretical_Chemistry)/Chemical_Bonding/Fundamentals_of_Chemical_Bonding/Chemical_Bonds/Bond_Lengths_and_Energies", "573876": "[Here](https://www.kaggle.com/soerendip/angle-and-dihedral-for-the-champs-structures) is the output of my [kernel](https://www.kaggle.com/soerendip/calculate-angles-and-dihedrals-with-networkx) for angle and dihedral calculations.\n", "559915": "supplemental data from here: https://pubs.acs.org/doi/suppl/10.1021/acs.jctc.7b00577/suppl_file/ct7b00577_si_001.pdf", "556824": "Will use QM9.\n\nDoes the usage of other open-source libraries have to be disclosed? I see that people are sharing links to packages on this thread, but is it compulsory to do so? To be more specific, I am referring to packages that do not explicitly contain or make use of quantum chemical data.", "556152": "Openbabel.", "556071": "How about the total energies of the involved atoms? Would it be allowed to calculate the energy of H, C, O, N, ... using e.g., GPAW?", "551741": "As of this day forward I might be using QM7 and QM9 datasets.", "549918": "https://github.com/samoturk/mol2vec", "548603": "Is this allowed \n\nhttps://www.kaggle.com/c/champs-scalar-coupling/discussion/95047#latest-548595?", "548323": "https://amp.readthedocs.io/en/latest/installation.html\n\nhttps://singroup.github.io/dscribe/install.html", "547578": "https://mendeleev.readthedocs.io - for some basic features of elements", "543872": "Does measured NMR spectra are allowed as external data? Given it is publicly available, open data.\n", "604783": "", "604516": "", "602887": "", "580194": "", "573877": "", "569604": "", "559591": "", "552365": "Can we use all available methods in openbabel? Or are there some specific methods that are not allowed?", "552117": "" } }