Extract experimental properties of any PubChem compound with ease.
Documentation:
Why, indeed, do we need another PubChem-related python package? The answer is that I've found the available one not quite fitting my needs. While the mentioned package offers a wide range of possible uses and allows to retrieve various computed parameters, it does not allow to retrieve any experimental ones. So I decided to write my own package that would do exactly that!
pip install pubchemprops
or
pip3 install pubchemprops
if you have several Python versions installed.
Please keep in mind that this package is written in Python 3, so, sadly, it won't work if you only have Python 2.x installed. Raise an issue if you'd like to see this kind of back-version support implemented, though, and we will see what we can do. :)
from pubchemprops.pubchemprops import X ,
where X is one of the following functions:
get_cid_by_name– takes a compound name, searches PubChem for it and returns it's PubChem IDget_first_layer_props– takes a compound name and a list of required parameters that CAN be retreived directly using the amazing PubChem PUG REST APIget_second_layer_props- takes a compound name and a list of required parameters that CAN NOT be retreived directly and for which one would have to look for in the depth of the whole PubChem record for the compound
This is a pretty simple task, but it may be useful if you would like to do something else with PubChem – that is, something this module cannot yet do.
Use it like that:
print(get_cid_by_name('acetone'))
The function accepts a compound's name (can be either IUPAC or rational) and returns its PubChem CID.
180
Just couldn't find any better name for that :] PRs much welcomed if you have another suggestion, though.
Basically what it means is that first layer properties are easy to retrieve, because there is a clear API for that. Here is the list of these properties:
MolecularFormula, MolecularWeight, CanonicalSMILES, IsomericSMILES, InChI, InChIKey, IUPACName, XLogP, ExactMass, MonoisotopicMass, TPSA, Complexity, Charge, HBondDonorCount, HBondAcceptorCount, RotatableBondCount, HeavyAtomCount, IsotopeAtomCount, AtomStereoCount, DefinedAtomStereoCount, UndefinedAtomStereoCount, BondStereoCount, DefinedBondStereoCount, UndefinedBondStereoCount, CovalentUnitCount, Volume3D, XStericQuadrupole3D, YStericQuadrupole3D, ZStericQuadrupole3D, FeatureCount3D, FeatureAcceptorCount3D, FeatureDonorCount3D, FeatureAnionCount3D, FeatureCationCount3D, FeatureRingCount3D, FeatureHydrophobeCount3D, ConformerModelRMSD3D, EffectiveRotorCount3D, ConformerCount3D, Fingerprint2D
Whoa, ain't that a hell of a lot of properties? And you can ask for any of those with the get_first_layer_props function! Use it like that:
easy_properties = get_first_layer_props('acetone', ['MolecularWeight', 'IUPACName', 'CanonicalSMILES', 'InChI'])
print(easy_properties) will return
{'CID': 180, 'MolecularWeight': 58.08, 'CanonicalSMILES': 'CC(=O)C', 'InChI': 'InChI=1S/C3H6O/c1-3(2)4/h1-2H3', 'IUPACName': 'propan-2-one'}
Okay, now moving on to the second layer. The name presumes that these properties are much harder to retrieve and you have to dig deeper to get to them. There is no direct API to acceess them or present them in a nice way. Still there are some pretty much interesting properties that you can get out of that:
IUPAC Name, InChI, InChI Key, Canonical SMILES, Wikipedia, Boiling Point, Melting Point, Flash Point, Solubility, Density, Vapor Density, Vapor Pressure, LogP, Stability, Auto-Ignition, Viscosity, Heat of Combustion, Heat of Vaporization, Surface Tension, Ionization Potential, Dissociation Constants
Also pretty big list, right? And you can get retrieve any property you like using the get_second_layer_props (provided there is a record for that property on PubChem itself). Example would be like following:
lysine_props = get_second_layer_props('L-lysine', ['IUPAC Name', 'Canonical SMILES', 'Boiling Point', 'Vapor Pressure', 'LogP'])
print(lysine_props) will return a dictionary like this:
{'IUPAC Name': [{'ReferenceNumber': 38, 'Name': 'IUPAC Name', 'StringValue': '(2S)-2,6-diaminohexanoic acid'}], 'Canonical SMILES': [{'ReferenceNumber': 38, 'Name': 'Canonical SMILES', 'StringValue': 'C(CCN)CC(C(=O)O)N'}], 'Vapor Pressure': [{'ReferenceNumber': 22, 'Name': 'Vapor Pressure', 'Description': '**PEER REVIEWED**', 'Reference': ['Daubert, T.E., R.P. Danner. Physical and Thermodynamic Properties of Pure Chemicals Data Compilation. Washington, D.C.: Taylor and Francis, 1989.'], 'StringValue': '5.28X10+9 mm Hg at 25 deg C /extrapolated/'}], 'LogP': [{'ReferenceNumber': 13, 'Name': 'LogP', 'Reference': ['HANSCH,C ET AL. (1995)'], 'NumValue': -3.05}, {'ReferenceNumber': 22, 'Name': 'LogP', 'Description': '**PEER REVIEWED**', 'Reference': ['Hansch, C., Leo, A., D. Hoekman. Exploring QSAR - Hydrophobic, Electronic, and Steric Constants. Washington, DC: American Chemical Society., 1995., p. 25'], 'StringValue': 'log Kow = -3.05'}, {'ReferenceNumber': 24, 'Name': 'LogP', 'Reference': ['HANSCH,C ET AL. (1995)'], 'StringValue': '-3.05'}]}
Looking messier than the first layer props return, but I didn't call 'em the second layer props for no reason, you see. :)
Still this is much better than having no info at all!
There is still a lot of work to be done:
- Add error handlers – oftentimes users will not provide a correct compound name due to typos or whatever, so we'd need to add some handlers for that. Ain't got none at the moment.
- Add more functionality – there are still lots of things one can retrieve from PubChem: images, spectra, bioinformation...
- Write better docs maybe
- Make the data returned look better and easier to read
- ???
- PROFIT!
PRs are very much welcomed, also feel free to open any issues or start discussions.
Hope you like the package!