pynmranalysis 1.1.3

python library for NMR preprocessing and analysis

Pynmranalysis

Python library for NMR preprocessing and analysis

Pynmranalysis has the ability to work with 1H NMR spectrum and offers many preprocessing functions that makes analysing the spectrum more effective Also it can be used to perform statistical modeling with great plots

• Preprocessing steps
• Normalization
• Statistical analysis

Installation

Install the pachage with pip command

pip install pynmranalysis

You may also install directly from this repository for the current master:

pip install git+git://github.com/1feres1/pynmranalysis.git

Dependencies : 'numpy == 1.20.3 ' , 'pandas == 1.2.4 ' ,'scipy == 1.6.3' ,'scikit-learn == 0.24.2' ,'matplotlib == 3.4.2'

Online Demo

The following python script shows you how to use the main functions of our library in this demo we will perform preprocessing steps on 1HNMR dataset then scale this data using NMR specific normalization function and finaly we will perform statistical analysis methodes like PCA and PLS-DA

demo link:

https://colab.research.google.com/drive/1A5qS1ObiiYBXmPnlecCTxzV41BzQ3fG6?usp=sharing

How to use

Preprocessing

A CSV file containing 1H-NMR spectra for 71 serum samples of patients with coronary heart disease (CHD) and healthy controls is located in CHD.csv in the exemple folder of this repository

# import 
import matplotlib.pyplot as plt
import pandas as pd
#read coronary heart disease data
spectrum = pd.read_csv("CHD.csv")
#convert columns from string to real numbers
columns = [float(x) for x in spectrum.columns]
spectrum.columns  = columns

Binning / Bucketing

In order to reduce the data dimensionality binning is commonly used. In binning the spectra are divided into bins (so called buckets) and the total area within each bin is calculated to represent the original spectrum

from pynmranalysis.nmrfunctions import binning
binned_data = binning(spectrum ,width=True ,  bin_size = 0.04 , int_meth='simps' , verbose=False)

fig , axs = plt.subplots(2,1 , figsize = (16,5))
fig.tight_layout()
axs[0].plot(spectrum.iloc[0] )
axs[0].set(title = 'spectrum before binning')
axs[1].plot(binned_data.iloc[0] )
axs[1].set(title = 'spectrum after binning')
plt.show()

Region Removal

By default, this step sets to zero spectral areas that are of no interest or have a sigificant and unwanted amount of variation (e.g. the water area).

from pynmranalysis.nmrfunctions import region_removal
r_spectrum = region_removal(spectrum=binned_data )

fig , axs = plt.subplots(2,1, figsize = (16,5))
fig.tight_layout()
axs[0].plot(binned_data.iloc[0] )
axs[0].set(title = 'spectrum before region removal')
axs[1].plot(r_spectrum.iloc[0] )
axs[1].set(title = 'spectrum after region removal')
plt.show()

Note : The implementation provided of those functions here is semilar to that of the R PepsNMR library [1].

Normalization

Mean Normalization

Each spectrum is divided by its mean so that its mean becomes 1.

from pynmranalysis.normalization import median_normalization
norm_spectrum = median_normalization(r_spectrum , verbose=False)

fig , axs = plt.subplots(2,1, figsize = (16,5))
fig.tight_layout()
axs[0].plot(r_spectrum.iloc[0] )
axs[0].set(title = 'spectrum before normalization')
axs[1].plot(norm_spectrum.iloc[0] )
axs[1].set(title = 'spectrum without normalization')
plt.show()

Median Normalization

Each spectrum is divided by its median so that its median becomes 1.

from pynmranalysis.normalization import quantile_normalization
norm_spectrum = quantile_normalization(r_spectrum , verbose=False)

fig , axs = plt.subplots(2,1, figsize = (16,5))
fig.tight_layout()
axs[0].plot(r_spectrum.iloc[0] )
axs[0].set(title = 'spectrum before normalization')
axs[1].plot(norm_spectrum.iloc[0] )
axs[1].set(title = 'spectrum without normalization')
plt.show()

Quantile Normalization

Each spectrum is divided by its first quartile so that its first quartile becomes 1.

from pynmranalysis.normalization import mean_normalization
norm_spectrum = mean_normalization(r_spectrum , verbose=False)

fig , axs = plt.subplots(2,1, figsize = (16,5))
fig.tight_layout()
axs[0].plot(r_spectrum.iloc[0] )
axs[0].set(title = 'spectrum before normalization')
axs[1].plot(norm_spectrum.iloc[0] )
axs[1].set(title = 'spectrum without normalization')
plt.show()

Peak Normalization

Each spectrum is divided by the value of the peak of the spectrum contained between "peak_range" inclusive (i.e. the maximum value of spectral intensities in that interval).

from pynmranalysis.normalization import peak_normalization
norm_spectrum = peak_normalization(r_spectrum , verbose=False)

fig , axs = plt.subplots(2,1, figsize = (16,5))
fig.tight_layout()
axs[0].plot(r_spectrum.iloc[0] )
axs[0].set(title = 'spectrum before normalization')
axs[1].plot(norm_spectrum.iloc[0] )
axs[1].set(title = 'spectrum without normalization')
plt.show()

PQN Normalization

Probabilistic Quotient Normalization from Dieterle et al. (2006). If ref.norm is "median" or "mean", will use the median or the mean spectrum as the reference spectrum ; if it is a single number, will use the spectrum located at that row in the spectral matrix; if ref.norm is a numeric vertor of length equal to the number of spectral variables, it defines manually the reference spectrum.

from pynmranalysis.normalization import PQN_normalization
norm_spectrum = PQN_normalization(r_spectrum , verbose=False)

fig , axs = plt.subplots(2,1, figsize = (16,5))
fig.tight_layout()
axs[0].plot(r_spectrum.iloc[0] )
axs[0].set(title = 'spectrum before normalization')
axs[1].plot(norm_spectrum.iloc[0] )
axs[1].set(title = 'spectrum without normalization')
plt.show()

Note : The implementation provided of those functions here is semilar to that of the R PepsNMR library [1].

statistical analysis

PCA

A pickle file containing 1H-NMR spectra for 64 serum samples of patients with two groups of disgstive diseases bliary/Pancreatic Disease and Intestinal Diseases is located in digestive_disease_data.pkl in the exemple folder of this repository

# import 
import matplotlib.pyplot as plt
import pandas as pd
#read data
data = pd.read_pickle('digestive_disease_data.pkl')
# split data into predictive variables (spectrums) and target varibles (digestive disease group)
# target -->  1 :Biliary/Pancreatic Diseases | 0 : Intestinal Diseases
spectrum = data.iloc[ : , :-1]
target = data.iloc[ : , -1].values

PyPCA

Principal component analysis, or PCA, is a statistical procedure that allows you to summarize the information content in large data tables by means of a smaller set of “summary indices” that can be more easily visualized and analyzed

from pynmranalysis.analysis import PyPCA
#create pypca instance 
pca = PyPCA(n_comps=3) 
#fit the model to data
pca.fit(spectrum)

Score plot is the projection of samples in the data set in lower dimention spce of the first 2 componants of the

pca.score_plot()

Scree plot is agraph that show each componant of the pca model with their explained variance

pca.scree_plot()

Outiler plot is a plot that calculate index of outliers in the data and plot them with different color

pca.outlier_plot()

Target plot is a scatter plot that shows the projection of each simple in the first 2 componants with Colors that much their classses in the target variable

pca.target_plot(target)

PyPLS_DA

Partial least squares-discriminant analysis (PLS-DA) is a versatile algorithm that can be used for predictive and descriptive modelling as well as for discriminative variable selection.

from pynmranalysis.analysis import PyPLS_DA
#create pyplsda instance 
plsda = PyPLS_DA(ncomps=3) 
#fit the model to data
plsda.fit(spectrum , target)

Interia plot is a paired barbot that shows R2Y (goodness of the fit ) score and R2Y (goodnes of predection with cross validation)

plsda.inertia_barplot(spectrum, target)

PLSDA score plot is a scatter plot that shows the projection of simples in the first 2 latent variables

plsda.score_plot(target)

Note : The implementation provided of those functions here is semilar to that of the R PepsNMR library [2].

License

Copyright (c) [2021] [Feres Sakouhi]

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

References

[1] PepsNMR for 1 H NMR metabolomic data pre-processing Manon Martin , Benoît Legat

[2] Partial least square for discrimination Matthew Barker1 and William Rayens