CheSweet: chemical shifts for glycans
The knowledge of the tridimensional structure of glycans is necessary to understand in detail, at atomic level, the molecular processes in which they are involved. Chemical Shifts (CS) are observables obtained from Nuclear Magnetic Resonance experiments that are highly sensitive probes to sense conformational changes. CS can be calculated accurately using quantum chemical tools, although these computations are very demanding for routine computations of more than a few conformations. For that reason we have developed CheSweet, a Python module for accurate and fast computation of CS.
CheSweet requires:
- Python (>=3.3)
- Numpy (>= 1.8.2)
- Scipy (>= 0.13.3)
This packages will be automatically installed if you don't have them yet.
-
Download this repository and unzip it.
-
Execute the following command in your terminal:
pip install chesweet-master/
-
Clone this repository
-
Execute the following command in your terminal:
pip install chesweet/
After installation is done you can run the automatic tests (optional), you will need to have pytest package installed, depending on the way that you downloaded the folder run:
cd chesweet-master
or
cd chesweet
And then run the test with:
pytest
This will check that all functions of CheSweet work in your installation. You can also use the nose package instead of pytest.
import chesweet as chswThe main argument of load is full. Passing False (default) is useful when working with the glycosidic torsional angles phi (Φ), psi (Ψ), and omega (Ω), True allows you to work with the mentioned torsional angles and also included the chi (Χ) torsionals. In the image below you can see the mentioned torsional angles for α-D-Glcp-(1-4)-α-D-Glcp.
The names of the disaccharides used by CheSweet follow the IUPAC syntax, but without the parentheses in the bond, e.g. for maltose the IUPAC name is a-D-Glcp-(1-4)-a-D-Glcp, CheSweet uses a-D-Glcp-1-4-a-D-Glcp.
Examples:
Load all calculated dissacharides in CheSweet:
# Reduced version of the datasets (using only phi, psi and omega)
disaccharides_red = chsw.CheSweet()
# Using full=True (using all torsionals)
disaccharides_full = chsw.CheSweet(full=True)You have the option of changing the location of the folder lut, in that case, you have to indicate the new location on the argument path.
Once you have loaded the look-up tables it is possible to calculate the chemical shifts or the torsional that you want.
In the next examples we show how to use these functions for maltose [α-D-Glcp-(1-4)-α-D-Glcp].
Using the reduced look-up table you have to provide the name of the disaccharide and the values of Φ and Ψ torsionals:
chemC1_red, chemCx_red = maltose_red.compute_cs('a-D-Glcp-1-4-a-D-Glcp', 85.3, 76.8)
print(chemC1_red, chemCx_red)
102.852047 76.693767When using the full look-up table you have to also provide the Χ torsional angles:
# Using full=True
chemC1_full, chemCx_full = maltose_full.compute_cs('a-D-Glcp-1-4-a-D-Glcp', 85.3, 76.8, 46.8, 78.6, 173.1)
print(chemC1_full, chemCx_full)
101.85525 72.19692You can see in these examples that depending on the datasets used (reduced or full) the obtained result can have little differences (Garay et. al 2014).
This example is the inverse of the previous one. We are now passing the CS for the carbons involved in the glycosidic bond and we are getting the compatible torsional angles given a tolerance eps, by default eps=0.5. The first CS should be C1 and the second CS should be the second carbon in the glycosidic bond.
# Reduced version of the datasets
tors_array_red = maltose_red.compute_tors('a-D-Glcp-1-4-a-D-Glcp', 109.52, 87.68)
print(tors_array_red)
[[ 90. 120.]
[ 110. 100.]
[ 130. 130.]
[ 140. 100.]]In this case, we obtain information only about Φ and Ψ, first and second column respectively.
You can obtain more values for these torsionals by increasing the value of the parameter eps (by default 0.5):
# augmented eps
tors_array_red2 = maltose_red.compute_tors('a-D-Glcp-1-4-a-D-Glcp', 109.52, 87.68, eps=0.6)
print(tors_array_red2)
[[ 90. 120.]
[ 100. 110.]
[ 110. 100.]
[ 130. 90.]
[ 130. 130.]
[ 140. 100.]]Alternatively, you can use the full datasets.
# Using full=True
tors_array_full = maltose_full.compute_tors('a-D-Glcp-1-4-a-D-Glcp', 109.52, 87.68)
print(tors_array_full)
[[ 80. 120. -60. 180. -60.]
[ 90. 110. 180. 180. -60.]
[ 90. 110. -60. 60. 60.]
[ 90. 120. -60. -60. -60.]
[ 90. 130. 180. 60. -60.]
[ 100. 100. 180. 180. -60.]
[ 100. 130. 60. 180. 180.]
[ 110. 90. 180. 180. -60.]
[ 110. 100. -60. -60. -60.]
[ 110. 130. 60. 180. 180.]
[ 110. 160. 60. 60. 60.]
[ 110. 160. 180. 60. 60.]
[ 120. 100. 60. 60. -60.]
[ 130. 100. 60. 60. -60.]
[ 130. 140. 180. 60. 60.]
[ 140. 130. -60. -60. -60.]]You can see in these examples that depending on the datasets used (reduced or full) the obtained result can have little differences and be using the full version of the dataset you have additional information about the Χ angles.
For each look-up table file (lut file, for short) the last two columns are the values of the pre-calculated CS. The first of that columns corresponds to the C1 and the last correspond to the second carbon in the glycosidic bond (Cx). The remaining columns are the torsional angles.
There are two types of lut files for each disaccharide, one type (called "full") includes all the torsional angles (lut files with the name of the disaccharide) and the other type (called "reduce") only includes the torsionals of the glycosidic bond (lut files with the name ended with "_red"). Which of this files is used in the calculation is controlled with the keyword full.
Last 4 lines in the file "a-D-Glcp-1-1-a-D-Glcp", the header with the columns names is not present in the actual lut files and is show here only to help identify the content of each column:
phi psi chi1 chi2 nan nan C1 CS C1' CS
-10 130 -60 -60 0 0 87.3016 87.6553
-10 140 60 60 0 0 85.9775 86.1810
-10 140 180 60 0 0 86.0382 85.3786
-10 140 -60 60 0 0 84.7857 85.0489
Here the chi1 belongs to the first residue and chi2 to the second residue
Last 4 lines in the file "a-D-Glcp-1-4-a-D-Glcp":
phi psi chi1 chi2 chi3 nan C1 CS C4' CS
-10 130 180 -60 -60 0 89.1629 117.2891
-10 130 -60 60 -60 0 88.2327 117.6398
-10 130 -60 180 -60 0 88.2642 115.4036
-10 130 -60 -60 -60 0 87.5465 116.9395
Here chi1 belongs to the first residue and chi2-chi3 to the second residue
Last 4 lines in the file "a-D-Glcp-1-6-a-D-Glcp":
phi psi chi1 chi2 chi3 nan C1 CS C6' CS
-10 -90 180 180 -60 0 85.2663 114.0324
-10 -90 180 -60 60 0 84.0807 113.5660
-10 -90 180 -60 180 0 84.9781 113.5554
-10 -90 180 -60 -60 0 83.9020 113.9225
The columns with all-zeros are not used in the calculations, these columns are just placeholders in case a dissacharide with an additional torsional angle is included in the future.
In contrast to the full lut files, that have a variable number of columns all reduced lut files have 4 columns, except for 1-6 glycosidic bonds were an omega angle included, as you can see below.
First 4 lines in the file "a-D-Glcp-1-4-a-D-Glcp_red":
phi psi C1 CS C4' CS
-10 100 89.6864 116.5005
-10 110 90.2366 114.8895
-10 120 89.9299 116.8816
-10 130 89.0696 116.8426
Last 4 lines in the file "a-D-Glcp-1-6-a-D-Glcp_red":
phi psi omega C1 CS C6' CS
-10 -110 180 85.9299 117.8878
-10 -100 60 84.3075 119.8173
-10 -100 180 85.5578 115.8875
-10 -90 180 85.0815 113.8175
If you need help in using this package or you found a bug please open an Issue.
All contributions are welcome!
You can contribute via GitHub, send your proposed changes with your Pull Requests and the errors/improvements/comments must be reported by Issues.
Garay, P. G., Martin, O. A., Scheraga, H. A., & Vila, J. A. (2014). Factors affecting the computation of the $^{13}$C shielding in disaccharides. Journal of Computational Chemistry, 35(25), 1854–1864. https://doi.org/10.1002/jcc.23697
This project is licensed under the MIT License - see the LICENSE file for details.