Pandas?

[kyleabeauchamp]

Hey, have you every considered using Pandas? It provides a lot of the database functionality that you're using. I think Pandas could be helpful for what you're trying to do--although I realize the learning curve is a bit steep.

[mattfenwick]

@kyleabeauchamp thanks for the suggestion, I will definitely be trying Pandas out today.

If you don't mind my asking, are you doing NMR or bioinformatics? I'm just curious how you came across this project.

Thanks again!

[kyleabeauchamp]

So I do MD simulation, but I deal with a lot of NMR data. I've used your PyStar reader.

On Tue, Aug 13, 2013 at 4:36 AM, Matt Fenwick notifications@github.comwrote:

@kyleabeauchamp https://github.com/kyleabeauchamp thanks for the suggestion, I will definitely be trying Pandas out today.

If you don't mind my asking, are you doing NMR or bioinformatics? I'm just curious how you came across this project.

Thanks again!

— Reply to this email directly or view it on GitHubhttps://github.com/mattfenwick/NMRPlay/issues/1#issuecomment-22558805 .

[kyleabeauchamp]

I've personally been using Pandas DataFrame objects to store Ensembles of chemical shifts--one dimension is the ensemble, the other dimension is the different atoms. I then use Pandas Series to store the measured values.

On Tue, Aug 13, 2013 at 9:07 AM, Kyle Beauchamp kyleabeauchamp@gmail.comwrote:

So I do MD simulation, but I deal with a lot of NMR data. I've used your PyStar reader.

On Tue, Aug 13, 2013 at 4:36 AM, Matt Fenwick notifications@github.comwrote:

@kyleabeauchamp https://github.com/kyleabeauchamp thanks for the suggestion, I will definitely be trying Pandas out today.

If you don't mind my asking, are you doing NMR or bioinformatics? I'm just curious how you came across this project.

Thanks again!

— Reply to this email directly or view it on GitHubhttps://github.com/mattfenwick/NMRPlay/issues/1#issuecomment-22558805 .

[kyleabeauchamp]

FYI: here's an example of something I've done with Trialanine--you can download FitEnsemble from my github:

from fitensemble import example_loader
predictions, measurements, uncertainties = example_loader.load_alanine_pandas()

In [8]: predictions.mean()
Out[8]: 
J3_HN_HA    6.592816
dtype: float64

[kyleabeauchamp]

We also have some (still in beta testing) ability to take a PDB or trajectory file and convert it into a pandas DataFrame. That is through our MDTraj library (https://github.com/rmcgibbo/mdtraj)

 >>> topology = md.load('example.pdb').topology
>>> print topology # doctest: +SKIP
<mdtraj.Topology with 1 chains, 3 residues, 22 atoms, 21 bonds at 0x105a98e90>
>>> table, bonds = topology.to_dataframe()
>>> print table.head()
serial name element resSeq resName chainID
0 0 H1 H 0 CYS 0
1 1 CH3 C 0 CYS 0
2 2 H2 H 0 CYS 0
3 3 H3 H 0 CYS 0
4 4 C C 0 CYS 0
>>> # rename residue "CYS" to "CYSS"
>>> table[table['residue'] == 'CYS']['residue'] = 'CYSS'
>>> print table.head()
serial name element resSeq resName chainID
0 0 H1 H 0 CYSS 0
1 1 CH3 C 0 CYSS 0
2 2 H2 H 0 CYSS 0
3 3 H3 H 0 CYSS 0
4 4 C C 0 CYSS 0